Griffiths

Looking back upon the argument as it unfolds in this article, it strikes me that its structure is essentially autobiographical. The effectiveness of legislation is a classic topic in the sociology of law, but one which at the time I began working in the field was already widely regarded as outmoded and wrongheaded (see Griffiths 1978). The article begins with that conclusion and an analysis of the fundamental reasons why the traditional, top-down ‘instrumentalist’ approach to legislation had proven sterile and, as I argue, unsociological and untenable. I then propose a bottom-up alternative - the ‘social working approach’ - which concentrates on rule following on the ‘shop floor’ of social life, the social organization of which is conceptualized in terms of ‘semi-autonomous social fields’. The greater part of the article is devoted to exploring the implications of such a theoretical choice. As the argument progresses, regulation by semi-autonomous social fields is seen largely to set the conditions under which actors follow ‘legal’ rules. And the analysis of rule following entails attention not only to primary rules of behavior but also to a whole complex of secondary rules governing the mobilization of primary rules. Rule-following behavior thus appears as the product of following rules about following rules. In other words, the effectiveness of law is socially regulated.

At various points in the argument the source and adaptedness to the local situation of a rule seem to be important. This idea leads to an exploration of the phenomenon of legal change originating on the shop floor itself: self-regulation. Semi-autonomous social fields are not only the social locus of rule following but also of the processes by which what ultimately become ‘legal’ rules emerge. And this local ‘legislative’ process is itself largely a matter of the application of rules. The continuities between ‘legal’ and ‘non-legal’ rules are thus more manifold and more profound than one had realized. And one of the few (latent) assumptions of instrumentalism which was regarded at the outset of the article as theoretically sound - that it is possible and necessary to make a neat theoretical distinction between rules as dependent and rules as independent variables - turns out to be ill conceived.

A quarter of a century after having come to the depressing conclusion that the sociology of law had not produced a single interesting, general proposition about legislative effectiveness, I now conclude optimistically. It does seem possible to give a general, theoretical, sociological answer to the question with which I began: Under what conditions do people follow a legal rule?

The following anecdote about the use of legal rules in everyday social life was told by my colleague Albert Klijn at a small colloquium devoted to an earlier version of this article.

Traveling first class by train from Utrecht (where I live) to Groningen (where I work), a trip of about two hours, I have four options in choosing a seat: smoking or non-smoking open carriages, six-person compartments where smoking and talking are allowed, and non-smoking six-person ‘work-compartments’ for those who want to work quietly (these are clearly marked as such on the door with an icon of a man with his fingers to his lips and the words ‘Silence. Work-coupé ’). I prefer the last option, since I am a non-smoker and always have a lot of work to do. Normally one enters the compartment, greets one’s fellow passengers, sits down to work, and nothing untoward happens during the trip.

But one can be disappointed. Sometimes other people sit in the compartment and start to talk. I usually do nothing until Amers[page 3]foort (a major train junction about 15 minutes into the trip), hoping they will get out there. I can read the morning paper and their conversation will not bother me much.

If they do get out at Amersfoort, I am happy. But if they turn out to be going in the direction of Groningen too, and go on talking after Amersfoort, I say something like “Gentlemen (or ladies), this compartment is for working, you know.” I might call their attention to the icon. Usually this is enough. One time they reacted by saying, “We are working: we are having a meeting!” But they agreed that they had to be silent and so they were.

Very rarely, one’s fellow passengers do not stop talking even when asked. After Amersfoort I really have to get some work done and I need the silence I am ‘entitled’ to. I have three choices: try to ignore the disturbance, look for another seat, or make them stop talking by threatening to complain to the conductor. Usually I have decided in advance that if they do not stop talking I will stick up for my rights. While they are sometimes rather unpleasant about it, I cannot recall a case in which threatening to invoke the authority of the conductor did not suffice.

Many of the elements of the theoretical approach to legal rules that I shall develop in this article are contained - if in embryonic, inexplicit form - in this anecdote from everyday life. One way to describe the purpose of the article is as that of trying to make good the lack of systematic theoretical analysis such situations have received. My claim is that until we understand the complicated ways in which legal rules are used in such everyday interaction, we will not be in a position to explain or predict much of real significance about the effects of legislation.

In thinking about the effects of legislation, two opposite sorts of bewonderment are possible. On the one hand, it is surprising that despite all the energy invested in them, legal rules often seem so ineffective. Attempts to influence behavior by subjecting it to rules, if studied carefully, generally turn out to have few desired [page 4] effects, or many undesired ones, and often both at the same time. If desired effects do occur, it is often unclear whether they were brought about by the rules in question.[3]

On the other hand, if one pauses for a moment to think about it, it is hard to understand how anyone could ever expect a legislated rule to have any effect on behavior. After all, as it leaves the legislative body a law seems to be nothing more than so many ink marks on paper. If legislation is more than this, exactly what that ‘more’ consists of needs to be specified and accounted for: the connection between a legislative text and actual behavior is not at all obvious. How it is that ink marks get transformed into regulated behavior, if and when they do, is one of the great mysteries of the sociology of law.[4] It is this mystery that I propose to tackle in this article.

The particular approach to that question which I present in this article I call the theory of the ‘social working’ of legal rules. Actually, in its present form it is mostly a sort of proto-theory, largely concerned with conceptual clarification and putting available knowledge and insight into coherent order. In short: a prolegomenon to the formulation of testable theoretical propositions.

I think of theory as generalized insight resulting from experience and useful for guiding intelligent action. In this article I present the results of a search for such insight into the social effects of legal rules, a search that in the first instance was intended to inform several specific research undertakings. The most important of these concerns the regulation of medical behavior that shortens life, an undertaking that initially was limited to euthanasia. The problem of regulating euthanasia has been in the forefront of my thinking and writing, empirical and otherwise, during the long gestation period in which the theory of the social working of legal rules slowly [page 5] took form. Important parts of the theory, as I now conceive it, first became clear to me in response to specific problems in the analysis of data concerning the regulation of euthanasia. A decade of work on the subject recently resulted in Euthanasia and Law in the Netherlands (Griffiths, Bood, and Weyers 1998), whose argument derives largely, if on the whole implicitly, from the social-working approach to legislation.[5]

A second research undertaking that helped form my thinking about the influence of legal rules on behavior had to do with the limitations and possibilities of law and legal administration in the realization of forest policy, specifically in connection with the conservation of the tropical forest in the Amazonian region of Ecuador.[6] Over the years I have also devoted both theoretical and empirical attention to the effectiveness of legal rules in combating discrimination and promoting emancipation, especially of women;[7] here, too, the social-working approach has played a key role.[8]

I mention these practical contexts here partly because I shall use them as examples throughout this article,[9] but more importantly because different sorts of legal rules presumably entail differences of theoretical emphasis. Although I try to think about the social working of legal rules from a general point of view, it seems likely that sorts of rules with which I have never been concerned may require additions and adjustments to the theory.

One important limitation is implicit in the nature of the research undertakings just described: they all concern primary rules of behavior. Not all legal rules are of this sort. Some rules provide facilities for people who wish to accomplish a particular objective (for example, many of the rules of contract law),[10] or incentives to behave in a particular way (such as subsidies and taxes), or procedures and structures within which activities take place (such as an adjudicatory system), or constitutional, interpretational and choice-of-rule rules about how to do things with primary rules of behavior. The theory presented here does not deal directly with these sorts of legal rules and is therefore not complete. I do claim, however, that it is fundamental, in the sense that understanding the social working of all other sorts of rules depends on understanding how primary rules of behavior work.

The limitation of much of the discussion to legislated rules[11] is merely a matter of convenience and the theory is ultimately far more ambitious. The social working of legislated rules is a special case of the social working of legal rules in general and that in turn is a special case of the social working of rules generally. The theoretical context I have in mind is thus that of a general theory of rule-following behavior. Since in my view (Griffiths 1984a) ‘social control’ is the “enterprise of subjecting human conduct to the governance of rules” - to borrow Fuller’s (1964: 91) expression for slightly different purposes - the objective I ultimately have in mind is a general theory of social control, or - to be supremely immodest - a theory of how social life is possible at all (assuming, as I do, that without rules[12] the life of man [page 7] would indeed be ‘solitary, poore, nasty, brutish and short,’ and in any event completely non-social).

The larger context in which the discussion is conceived permits me to postpone to another place consideration of two questions. One concerns the possible implications of the fact that legislation is the specific type of rule being considered: I shall pay no particular attention to the possible peculiarities of legislated rules of behavior as against, for example, rules whose source is customary or judicial. Nor, for that matter, will the possibility that among all rules there is something special about those that are ‘legal’ in the everyday sense of the word be explored.[13]

The second question I shall not address concerns the most fundamental and most obscure idea in a theory of social control: the notion of ‘following a rule’. I regard it as obvious that human beings have the capacity to follow rules and that the exercise of this capacity is often necessary in explaining individual behavior and always essential in accounting for social practices and institutions. The evidence for these assumptions is staring us in the face, mine as I write these words and yours as reader of this well-formed English sentence. Language is quintessentially an affair of rules and rule following. However little we may understand how it is possible, we do seem in practice to be able to communicate. I shall simply assume for present purposes that the notion of following a rule is conceptually coherent and empirically operationalizable, and that its features are more or less those that common sense ascribes to it.[14]

Although the question it deals with is thus a ‘big’ one, the pretensions of this article [page 8] are modest. There is not much that is really ‘new’ to be found in it. My main objective has been to bring together and present in a systematic way a number of ideas that have emerged in the literature in recent decades and that seem to me to afford the basis for a promising theoretical approach to legislation. I have tried to make explicit and coherent what has so far been latent and diffuse. If some readers conclude from this that the social-working perspective is ‘only’ a sophisticated form of the received approach to legislation (what I shall call ‘instrumentalism’), I have no problem, although it does seem to me a bit like saying that Copernican cosmology is ‘only’ a more sophisticated version of Ptolemaic cosmology. Of course, if your interest is immediately practical, you may be able to get from one place to another with either one. What is scientifically (and in the long run practically) important, it seems to me, is how you do it.

Before plunging into the substance of the argument, I need to draw some distinctions between different sorts of things that can be regarded as social effects of a legal rule. Such effects can be ‘direct’ or ‘indirect’, and direct effects can be ‘primary’ or ‘secondary’.[15] Effects can also be ‘special’ or ‘general’. And they can take place in [page 9] the context of a ‘trouble case’ or of a ‘trouble-less’ case. These distinctions are of key importance for the theory of social working.

In analyzing the social effects of a rule it is important to distinguish between the behavior provided for in the rule, which, when it occurs and the rule is responsible for it, constitutes the rule’s ‘direct effects’, and the social consequences of this behavior: the rule’s ‘indirect effects’.

The principal direct effect of a rule requiring drivers to keep to the right side of the road is right-side driving. Direct effects, in other words, consist of rule-following behavior. Literal conformity, however, is not the only way that a rule can be said to be followed. People may adapt their behavior to the existence of a rule even when they are breaking it, as in the case of a bicyclist paying extra careful attention while biking the wrong way on a one-way street (a common occurrence in Groningen, where I live and work). Avoidance is another possible direct behavioral effect: a rule (such as tax or marriage rules) that makes a particular sort of behavior especially costly may induce people to choose alternative sorts of behavior that are not subject to the rule. The lowest common denominator of all direct effects of a rule is that the rule is directly responsible for the behavior that is said to follow it.[16]

Indirect effects, on the other hand, are the social consequences of direct effects, of rule-following behavior. The principal indirect effects of a rule requiring right-side driving are better traffic coordination and a lower accident rate than there would be if there were no rule on the matter. It is usually indirect effects that are the real objective of legislative effort, direct effects in themselves being mostly a matter of indifference: even the worst sort of behavior is principally reprehensible because of [page 10] its indirect effects (for example, on its victims). It is for this reason that, as we will see shortly, the received ‘instrumentalist’ approach to legislative effectiveness concentrates on indirect effects. But, as we will also see, no general theory of the indirect effects of legislation is possible, which is why the social-working approach is limited to direct effects.

Among direct effects a further distinction must be made: between ‘primary’ and ‘secondary’ direct effects. A direct effect is primary when the behavior concerned implements the substantive behavioral standard contained in the rule (as I have noted above, this need not entail literal conformity). It is secondary when the rule following consists of efforts by one person to secure conformity with the rule by another.[17] A layperson can in this latter way ‘mobilize’ a rule by confronting a potential offender himself (as when Klijn reminds his fellow passengers that talking is prohibited in the work-compartment of a train, perhaps simply by pointing to the icon). He may do this individually or with the support of a group, and it may be on his own behalf or he may be seeking to maintain a general standard of behavior. Secondary mobilization of ‘legal’ rules in this ‘informal’ way is extremely common in social life, and it can be very subtle.[18]

But Klijn might have preferred, or been driven, to call the situation of (threatened) violation to the attention of an official charged with securing compliance, in this case the conductor. The enforcement behavior of such representatives of the modern state involves secondary rule-following that can be either ‘reactive’, as in Klijn’s example, or - much more rarely - on their own initiative: ‘proactive’.[19]

In the analysis of direct effects a distinction must further be made between the ‘special’ and the ‘general’ effects that the secondary mobilization of a rule can bring about. Special effects are effects in cases in which the rule is applied to a particular case, whether by an individual (calling attention to the ‘work-compartment’ icon) or by an official (such as a conductor who orders the offender to comply), or by a judge (imposing a sanction for failure to comply). The special effects of a rule are the primary rule following accomplished by such secondary mobilization. Secondary mobilization ‑ especially that done by ordinary people in the course of everyday interaction – can often be an important factor in the production of special effects. By calling attention to the ‘work-compartment’ icon, for example, one can usually get one’s fellow traveler to behave properly.

The ‘general’ effects of an instance of secondary mobilization of a rule are its effects in other situations, situations in which no secondary mobilization takes place. If my fellow-travelers refrain from talking in a ‘work-compartment’ without my having to call their attention to the icon, because they have witnessed (or heard about) cases of other people being taken to task for talking[20] or because this happened to them on some previous occasion, then we have a situation of ‘general effects’. In many circumstances the general effects of secondary mobilization - especially that by officials - are sociologically (and practically) far more important than the special effects.

Direct effects can be realized in the context of a ‘trouble case’ or in that of a ‘trouble-less case’. In a ‘trouble case’ one actor makes an appeal to a rule, but this is at least initially resisted. Most of the time, in such a case, the actors will be able to reach an accommodation. Only if they fail in this do they have a ‘dispute’ in which third parties may become involved. The emergence and life histories of disputes has in recent decades dominated the literature in sociology and anthropology of law (partly for the simple methodological reason that in the context of a dispute rules are much more readily visible than they are in trouble-less everyday social interaction). [page 12] But as Holleman (1973) argued, exclusive focus on ‘trouble cases’ tends to obscure the fact that by far the greater part of rule following takes place in situations in which actors carry on their affairs without any sort of dispute or dispute-settlement.[21] Rule following typically takes place as part and parcel of - and therefore not differentiated from - unproblematic social interaction. And as Galanter (1981) has argued, the most important social effects of a ‘trouble case’ often lie, not in the particular resolution it receives (special effects) but the influence of information about that resolution on the way other actors arrange their affairs and deal with (potential) disputes (general effects).

It follows from these considerations that if it is limited to trouble cases, theorizing concerning the social effects of rules will inevitably miss the greater part of the point.

The object of the theory of social working - what it seeks to explain - is the direct effects of rules. The theory focuses on the conditions under which and the processes by which the relevant actors ‑ those to whom a rule is primarily addressed, others in their immediate surroundings, officials ‑ follow or apply or enforce the rule. Of particular importance is rule following in trouble-less cases, which do not give rise to disputes (let alone disputes that come to the attention of officials). The special effects of informal mobilization can often be an important force for primary compliance, but the special effects of official enforcement are usually of limited social importance. These assertions are illustrated, for example, by the case of the no-talking icon. The essential regulatory problem here is, as it almost always is, to secure a high level of trouble-less rule following. Would-be talkers must apply the rule to their own behavior, and if they do not, the necessary secondary mobilization must come from their immediate social surroundings. Official mobilization of the rule is, and necessarily usually will be, rare, and mostly interesting for its general effects. If we can understand why the rule against talking in a ‘work-compartment’ is generally so effective,[22] we will have a grip on the fundamentals of effective regulation generally.

It is possible to approach the question of the effects of legislation in two paradigmatically different ways. Like all attempts to capture the endless variety of human behavior in a simple typology, each of them considered in itself as a description of the whole thought of any particular person would usually be little more than a caricature. The point is not to allocate individual social scientists to the one or the other paradigm but to identify and assess divergent tendencies of thought.

The familiar approach to the question of legislative effect is ‘instrumentalism’. Instrumentalism considers a legislated rule as a tool in the hands of a policy-maker who is bent on realizing (or forestalling) some sort of social change. Aubert’s (1966) classic study of the effectiveness of the Norwegian law on the rights of housemaids is an early and still one of the best studies within the instrumentalist paradigm, sensitive to many of its weaknesses.

The relationship between a rule and its social effects is conceived of in the instrumentalist paradigm as a straightforward causal one.[24] This conception (generally left implicit) has seriously inhibited the development of a theory of legislation for a number of reasons: (1) On the whole, instrumentalists exhibit little interest in the complexities of the supposed causal mechanism; the essential and perplexing question, ‘What does the possibility that a legal rule influences behavior presuppose?’, is scarcely asked. (2) Because instrumentalism is fundamentally a policy-maker’s point of view, it pays little attention to all those possible social effects of legislation that do not appear directly ‘relevant’ to the policy-maker’s (presumed) intent. (3) Tools are made to work, so if legislation is considered as a tool it is easy [page 14] to assume that it generally produces the desired results. The ‘normal’ state of affairs is therefore one of effectiveness and instrumentalists have primarily sought explanations for the supposedly ‘deviant’ situation of ineffectiveness.[25] (4) Rules are therefore treated as if they can be considered descriptions of actual behavior.[26]

Those who have seriously considered the matter concluded long ago that instrumentalism is a sterile approach to legislation, one that cannot lead to the development of significant theory.[27] Years of instrumentalist effectiveness research had led only to the monotonous conclusion that legal rules, so considered, rarely produce the intended results and are therefore apparently of little social significance. [page 15] With the notable exception of Black (1976), however, few have been daring enough to draw such a conclusion in public. For want of a better theoretical paradigm for thinking about the social effects of legal rules, sociology of law largely turned its back on legislation and devoted its attention instead to things like dispute processing and the legal profession. The result of all this is that the sociology of law has yet to produce a useable theory of legislation.

The instrumentalist paradigm has not only proven sterile in practice, the legal and social theory it implies cannot withstand examination. A legal rule is seen within this paradigm as a command issued by the legislator, addressed to individuals who are presumed to adjust their behavior so as to bring it into conformity with the command. Instrumentalism thereby makes a number of untenable assumptions concerning social organization and social life and the place therein of legal rules and institutions.[28] The most important for our purposes are these:[29]

1 The assumption of atomistic individualism. Society is seen as made up of individuals bound together by the state organization and not, essentially, by anything else. A theory of legislative effects concerns itself with the behavior of individual ‘rational actors’.

2 The assumption of perfect legal knowledge. The state organization is seen as a chain of command that transmits the commands of the legislator in uniform, undistorted form. And the social space between the state and the individual is conceived of as a normative vacuum through which the commands of the legislator pass unmediated and untransformed by intervening social rules and structures on their way to the individual. Instrumentalism treats the legally correct interpretation of the law ‑ the interpretation supposedly contemplated by the legislator ‑ as the command that reaches the individual and influences his behavior.

3 The assumption of legal monism. The state is assumed to have an effective monopoly over the regulation of interaction that (except in some extremely deviant situations such as the mafia) excludes other sources of regulation [page 16] as important influences on behavior.[30] The instrumentalist tradition in research looks not to competitive sources of regulation but to the recalcitrant self-interest or ‘deviant’ character of the individual as the prime suspect in case of non-compliance.

4 The assumption of legislative autonomy. The legislator is treated as external to and independent from the social context in which legal rules are effective. The idea of law as an instrument of social change presupposes this sort of Archimedean[31] legislative autonomy.

What is missing in this peculiar vision? Answer: attention to the real social state of affairs. Instead one finds a number of ideological preconceptions about how a society ought to be organized masquerading as a description of how it actually is organized.[32]

There are those who will object to the above sketch of instrumentalism, protesting that it is a straw man, that ‘nobody believes such nonsense anymore’. It is hard to refute such a claim by citing chapter and verse, precisely because instrumentalism, while in practice pervasive among legal policy-makers and socio-legal researchers (especially those who do research intended for the ears of policy-makers), is almost never made explicit. Most of its adherents appear quite unaware of the peculiar assumptions about law and society that form part of their everyday intellectual apparatus and would indignantly reject an effort to attribute these to them. A new approach must overcome the dead hand of unarticulated assumptions that in fact dominate public legislative practice and social science research but that, made explicit, nobody with his wits about him would want to defend. The essential [page 17] function of the ‘straw man’ argument is to permit those invoking it to relapse into their dogmatic slumber, to continue treating the way legislation supposedly has social effects as obvious and therefore not worthy of serious thought.

The key to a more adequate analysis of legislative effect can be given in the form of four propositions that are simply the opposites of the basic assumptions of the instrumentalist paradigm:

1 The assumption of the fundamentally social character of man. Both the people who together compose a society and all their individual and collective activities are ‘social through and through’.[33] Oppositions between ‘individual’ and ‘society’, between individual ‘preferences’ and social rules, between ‘self-control’ and ‘social control’, are false oppositions. Legal rules are addressed to social beings acting in a specific social context, not to asocial ‘rational actors’ seeking to ‘maximize their preferences’.[34]

2 The assumption of the socially contingent character of legal knowledge. Communication of legal information is always problematic. In the first place, the internal social organization of the state does not in practice look much like a chain of command. A message about the law that other organs of the state (bureaucracies, courts, police) transmit is generally not quite the same as ‑ and often quite different from ‑ the message that emerged from the legislative process. This limited and distorted message usually does not reach the relevant actors directly but is transmitted by various intermediaries that have limited capacities and resources and generally also have their own axes to grind (lawyers, professional organizations, interest groups, trade unions, political leaders, the press, etc.). The transmission process is, in other words, a ‘transformation’ process in which the message gets simplified and otherwise distorted and enriched with all sorts of additional information (e.g. concerning the various risks and costs associated with [page 18] following or nor following it[35]). The message about the law that ultimately comes to an actor’s attention ‑ if any message gets through at all ‑ is thus seldom the same as what the legislator ‘intended’. The socially mediated process of interpretation of a rule of which one has learned adds yet another occasion for transformation before a rule has an opportunity to be followed.

3 The assumption of legal pluralism. The state is but one of many sources of regulation. For an individual engaged in social interaction, the behavioral expectations of the state are frequently less well known, less clear, and in any case less pressing than those of other sources of regulation that are closer to the scene.

4 The assumption of the inseparability of legislation from social life. Legislation in a modern state is a highly differentiated social process: it takes place in a special way and is largely delegated to specialists. But it is definitely not an autonomous social process: the people who engage in it are participants in social life as a whole and, more importantly, the process itself is only one form of social interaction among many and in a constant state of mutual interaction with all the rest. The content of legislation is ultimately determined by the same factors that are affecting all other forms of social action. Legislation is an integral part of processes of ordering, conservation, and change in society; it is not a distinct and autonomous force acting on those processes.[36]

The above propositions come down to the following: The influence of legislation on behavior can only be understood in terms of social organization. An adequate theory of legislation must be based on a realistic, not an ideological model of the social situation and the social processes concerned.[37]

The contours of a sociological approach to the effects of legislation emerge from the foregoing discussion. The instrumentalist question ‑ when does a given legislated rule cause the intended results? ‑ is no longer paradigmatic. Instead, the interesting question is what difference the law makes in a concrete situation of behavioral choice. Not: Do individuals obey? but: What do people do? What exactly happens when this factor is added to or subtracted from the social equation? Attention is directed, in other words, to the social working of law.

Looking at legislation in terms of its social working approaches legislation not top-down, as is the case with instrumentalism, but bottom-up. It focuses attention on the ‘shop floor’ of social life: that concrete social situation where the social action and interaction that are the subject of regulation take place. If we take account of this concrete situation, in all its multifariousness and its complex, nitty-gritty reality, the question of ‘effectiveness’ appears in a new and more complex form: How much importance can we suppose that a given legal rule has for the activities of the actors whose behavior interests us? Do they know about it and how do they interpret it? Does the social control system to which the rule belongs have any practical jurisdiction over the situation? How does the rule, as the local actors know it, fit into the whole spectrum of considerations - practical exigencies and the demands of [page 20] competing rules - of which they must take account?

Instrumentalism, as we have seen, takes as its point of departure the intent of the legislator, and this intent is generally focused on indirect effects: the ultimate policy objective a rule is supposed to accomplish. But what instrumentalist researchers usually do in practice is limit themselves to the study of direct effects. They simply (and usually without argument) assume that the intended indirect effects are produced by the direct effects. Even Aubert (1966) only studied the extent to which the behavior of household servants and housewives was influenced by the Norwegian law on household personnel. But from the point of view of the legislator, such direct effects are intrinsically uninteresting. A maximum working day, vacations and the like are only really interesting in relation to the health, happiness, and the like of those who benefit from them. In short, instrumentalist research usually fails to deliver the goods on the question it takes to be central: does this tool work?

The failure in practice of the instrumentalist project is not merely a consequence of methodological and other practical difficulties in carrying out research. It is a theoretical failure. As a theoretical endeavor, instrumentalism is incoherent because a general theory of the relationship between rules and their indirect effects is inconceivable. The reason for this is simple enough.[38] No single theory could [page 21] possibly deal in general terms with the indirect effects of legislation since the only thing all indirect effects have in common is that they presuppose direct effects: rule following. The relationship (if there is one) between direct effects (such as keeping to the right) and possible indirect effects (such as a lower accident rate) is different for every different subject of regulation and is not a legal or sociological matter but one of some mixture of psychology, engineering, economics, physics and chemistry, and so forth. The instrumentalist project thus requires two sorts of theory: a theory of rule following and an infinite number of ‘policy theories’ which tie direct effects to indirect effects for every possible subject of legislation: one theory for the relationship between drinking and driving safety, another for the relationship between tenure arrangements and the sustainable use of agricultural land or fishery resources,[39] another for the relationship between the reporting requirement for euthanasia and the effectiveness of societal control over such decisions,[40] yet another for the relationship between legal form and the functioning of state enterprises,[41] etc. ad infinitum. Since no general theory of indirect effects is possible, it is not surprising that the instrumentalist tradition has never produced one.

On the other hand, a general theory of direct effects - of rule following - does seem conceivable. The assumption underlying the social-working approach is that the fundamental elements of rule following are the same for all different sorts of rules and for rules dealing with all different sorts of subjects. It is with the fundamental elements of rule following that the theory is concerned.

In section 3 we have seen that it is with the shop floor of social life that we need to begin in studying the effects of legislation. The first question the theory of social working must address is therefore this: What - in general, theoretical terms - does the social organization of the shop floor look like?

The social organization of the shop floor can best be described and analyzed in terms of the concept of a ‘semi-autonomous social field’ (SASF). The development of a sociological theory of legislation begins with Moore’s (1973) ‘discovery’ of the SASF.

Moore’s article is the culmination of a long, deviant, ‘legal pluralist’ tradition in sociology of law, one that emphasizes the primacy of ‘folk law’ and ‘indigenous social ordering’ over legislation and formal legal ordering as influences on social behavior. Sumner (1906) proclaimed that it is impossible to change ‘folk-ways’ with law. Ehrlich (1913) argued that the “living law” ‑ the law that actually regulates behavior ‑ is the “rules of behavior” that constitute the “inner order” of the “social associations” that make up society, and that those rules of behavior are different from the “rules for decision” used in communication among lawyers and regarded by them as “the law”. Fuller (1969) described law as a “language of interaction”, thereby emphasizing the way law provides the structure and the vocabulary that people use in everyday life to order and adjust their interaction. Law, he argued, is far more than an instrument of regulation or a dispute-processing institution; these are secondary, not primary functions of law. And the ‘language of interaction’ is in the first place an indigenous one: the primary ‘linguistic’ community is society itself.[42]

Although it did bring forth some striking insights, this earlier legal pluralist tradition in sociology of law never led to a coherent theory of legislation. Sumner, whose concern with legislation is closest to mine, offers no account of the social organization within which ‘folk-ways’ emerge and are maintained and simply denies in [page 23] general terms the capacity of law to change them: he does not analyze how a legal rule does or does not work. Ehrlich (1936), whose treatment of the social organization of non-state law is most useful for my purposes,[43] is ultimately concerned with a prescriptive theory of legal reasoning rather than a descriptive and explanatory theory of the social effects of legislation. Fuller’s emphasis on the use of law in trouble-less daily interaction stresses, as I do here, the social working of law rather than social engineering with law, but he does not discuss what happens when a legal rule differs from indigenous rules of daily interaction.

What Moore does is to take a number of fairly well known if diffuse ideas in this legal pluralist tradition and fuse them into a single clearly defined concept suitable for describing and analyzing the social working of legal rules. Her concept of a SASF focuses attention on the social context in which any legal rule must work, emphasizes the normative heterogeneity of that context, and indicates in rough terms what aspects of that context are important for the social working of legal rules and why they are important.

Moore defines a SASF in what she calls ‘functional’ terms: a SASF brings forth and maintains behavioral rules. The defining criterion that distinguishes a mere collection of individuals from a group in the sociological sense - a SASF - is that a SASF has members and (to some extent) regulates their behavior. Some SASFs are what Moore calls ‘corporate groups’, possessing the capacity for managing collective resources, undertaking group action, and carrying on external affairs. This is not true, however, for most SASFs, including the New York garment industry, which affords one of her examples.[44] However, such a more amorphous social field does bring forth rules of behavior and enforce them, and thereby orders its internal and external affairs.

If we conceive of social control as the ‘enterprise of subjecting human conduct to the governance of rules’, the SASF is the fundamental locus of social control. It is the social context within which socialization takes place: learning how to follow rules and which rules to follow and when.[45] It is where a rule of behavior ‘exists’: when the question is asked, ‘whose rule is this?’ the answer will always refer to some SASF. It is the place rules emerge and are elaborated, where they are used and if [page 24] necessary enforced. Thus, in addition to the state, SASFs such as a group-practice of doctors, a hospital ward, a village, a group of neighbors, a work unit in a factory, a business or other organization, the occupants of an apartment-building, a branch of industry, an educational institution, a prison or other ‘total institution’, a religious, immigrant or ethnic group, and so forth, all regulate the interactions of their members, albeit to varying extents.[46]

As a locus of regulatory activity, a SASF is only partially autonomous. It can regulate its internal affairs to a certain extent ‑ maintain its own rules and resist (more precisely, as we shall see: regulate) the penetration of competing external rules ‑ but its members are also members of many other social fields and as such exposed to many other sources of regulation. Thus, under normal conditions, the state ‑ one of the most prominent and for many purposes the most inclusive of the SASFs in a modern society[47] ‑ substantially constrains the regulatory autonomy of other SASFs. The autonomy of the state is of course also limited by other SASFs, as the reader of political news in the daily newspaper will be aware.[48]

In a society of any size a vast number of SASFs can be identified, varying from formal and formidable ‑ and even multinational ‑ corporate groups such as the [page 25] Catholic Church, to ephemeral and feeble loci of regulation such as (conceivably) a group of passengers on a cruise ship.

Despite all this variation, however, the only thing that is essential if a collection of people is to be regarded as a SASF is that it exhibit some degree of autonomous regulation of its members’ behavior. My colleague Klijn’s fellow-passengers in a ‘work-compartment’ of a train probably do not usually constitute a group with its own rules in any significant sense. Similarly, large and amorphous categories of people, such as smokers and non-smokers,[49] men and women, blacks and whites, the rich and the poor, and so forth, are usually not SASFs. (As the examples make plain, however, under special circumstances, group forming based on such characteristics does take place.)

In every social setting more complex than Robinson Crusoe’s, people are members of all sorts of SASFs, large and small, more or less autonomous, more or less exclusive or overlapping, all of them continuously enforcing (so far as their limited capacities permit) their behavioral expectations. It follows that, again excepting Robinson Crusoe, human beings are typically subject to a myriad of conflicting normative demands on their behavior. Figure 1 depicts the model of social organization implied by the concept of a SASF.

Depending upon the purposes and point of view of the observer, an essentially indeterminate number and variety of SASFs, in all sorts of relationships to each other and to an encompassing larger society, can be identified in most social settings. The fact that an attempt to count the SASFs in a given setting would be a fruitless exercise does not entail that the concept itself is indefinite. Given Moore’s definition, it is in principle always possible to answer the question whether for the purposes at hand any given collection of people has enough regulatory capacity to be regarded as a SASF. The concept ‘SASF’ (or ‘group’) does not differ in this respect from other concepts such as ‘structure’ or ‘rule’. As with any descriptive concept, the proof of the pudding is in the eating. The concept of a semi-autonomous social field has in fact proven itself in empirical research concerning social regulation by writers as otherwise different as Ellickson (1991: ‘close-knit group’[50]), Edelman (n.d.: ‘socio-legal field’), and Homans (1951: ‘group’). The advantage of the concept SASF over most other formulations of the same idea is that it emphasizes the (continuously variable) factor autonomy.

SASFs often restrict the imposition of ‘external’ law and institutions upon internal relationships.[51] Some SASFs so dominate the flow of information, inward to their members and outward about their members’ activities, that the first prerequisites for external control are missing. Others provide such attractive alternatives to external law that their members have no reason to invoke it, or they bring such pressure to bear against the invocation of external law that their members, dependent as they are on the continuing relationships that such SASFs control, wanting ‑ in Moore’s phrase ‑ to “stay in the game and prosper” (1973: 729), cannot afford the risk of flouting local expectations.[52]

Regulating behavior on the shop floor with external law entails successful intervention in relationships and interactions that may be of great importance to the participants and often are already locally regulated. The state is often unable to bring anything like as much regulatory force to bear as other SASFs that are closer to the scene. From the point of view of the actor on the shop floor, the behavioral expecta[page 27]tions of one or more local SASFs are often far more immediate and compelling than those of the state. The situation of a woman who is a potential beneficiary of an affirmative action program at the workplace is typical. She is intermittently exposed to the behavioral expectations of emancipatory legislation, but also and on a daily basis to those of her colleagues at the workplace itself, of her household and wider family, of her feminist friends (for instance, at the ‘women’s café ’ she frequents), of her trade union, of her church, and so forth. One or another of these may persuade her not to apply for a position, which, had she applied, she might - thanks perhaps to the affirmative action program - have obtained. Thus can the purposes of one source of regulation be frustrated by the countervailing expectations of others.[53] In short: the pervasive normative pluralism characteristic of social life is one of the main factors responsible for the fact that legislation often does not produce the intended results.

The regulatory autonomy of a SASF ‑ the degree to which, in a social context of competing rules, it can make its own rules ‘stick’ ‑ varies widely. Some SASFs, as the history of the Soviet Union’s efforts to remake society according to a uniform plan illustrate, manage to resist the full power of the state over long periods[54] while others - a group of neighbors in a modern city who fashion their own rule about where people may park their cars - owe their ephemeral autonomy largely to the fact that no one else particularly cares. Schwartz (1954) emphasizes features of the shop floor that contribute to its regulatory autonomy by ensuring that group members are quickly and accurately informed of each other’s behavior and are in a position to react to it favorably or unfavorably. Writers otherwise as diverse as Schwartz, Moore (1973), Ellickson (1991) and Homans (1951) emphasize the extent to which the members of a SASF are dependent upon the relationships that it regulates for access to resources that are important to them and for which they have no other source. In sum, the more a SASF approaches the situation of a ‘primary group’ (regular, ‘face-to-face’ contact), the greater the equality of its members, and the [page 28] more they are dependent on each other for vital resources, the more compelling and exclusive its internal social control will be. These factors explain, for example, the well-known fact that the economic production unit ‑ from factory to university department - is generally a tough, quite autonomous SASF, capable of resisting not only the law of the state but also external regulation by management.[55] Similarly, the high level of autonomy of SASFs in the medical sector ‑ partnerships of specialists, hospital wards, general and specialist professional associations - is well known to be responsible for the relative inability of all sorts of external rules to secure effective penetration to the medical shop floor.[56]

Up to this point I have followed Moore in regarding a SASF as the source of primary rules of behavior and of the social control capacity required to enforce them and to resist the local enforcement of the conflicting rules of other SASFs. Moore does not emphasize the necessary corollary, that secondary rules[57] ‑ rules about rules - also have their fundamental social locus in SASFs. Some of these secondary rules concern the internal affairs of a SASF, addressing questions such as how to determine what the primary rules are and how to apply and to change them; in sociological terms, it is these rules that provide for differentiated social control in groups that have it. Other, externally-oriented secondary rules concern the proper relationships between the SASF and other social fields, in particular the state and its law. We might call these latter secondary rules of a SASF its ‘sovereignty’ and ‘choice-of-rule’ rules, determining respectively which sorts of regulatory issues belong exclusively to the competence of the group itself and which rule is applicable when more than one SASF has a rule applicable to a given case.[58] It is not their primary rules as such but the secondary rules of other SASFs that influence the [page 29] social working of legal rules on the shop floor of social life.[59]

To illustrate the complexity of the normative situation that legal rules often encounter on the shop floor,[60] let me use the example of the legal regulation of euthanasia in the Netherlands (for the bare bones of the legal situation see the box on the next page).

Legislation may sometimes deal with situations that are otherwise completely unregulated and can therefore be usefully analyzed as an effort to influence ‘rational actors’ pursuing their individual self-interest, although it is hard to think of a convincing example. But in most cases such a perspective is quite misleading (unless one defines ‘self-interest’ so broadly that it covers every conceivable motive for behavior, in which case the ‘rational-actor’ approach becomes tautological). It seems in any case clear that no analysis of the regulatory problems surrounding the euthanasia practice of doctors will get very far unless it takes account of the fact that their behavior is predominantly moral behavior, that is, required by the social rules of the specific social contexts within which medical practice takes place. Doctors do what they do to an important extent because they have learned to follow all sorts of social rules, in particular those of their profession (and of more local SASFs within it, down to the level of a hospital ward). Even the most virulent critics of current Dutch euthanasia practice have not argued that the ‘self-interest’ of doctors is a serious part of the phenomenon.

Let us look more closely at the shop-floor rules that doctors follow in connection with euthanasia.

The rules of the medical shop floor may require of a doctor behavior different from that required by legal rules. There seems, for example, to be a medical rule requiring the alleviation of suffering of those in one’s care. This is presumably a professionalized version of the more general social rule requiring beneficence, with a more specific content and a more imperative character.[62] Many Dutch doctors [page 31] interpret this rule as requiring them to accede to the patient’s wish for euthanasia in an appropriate case. Thus doctors do what they do in cases of euthanasia, to an important extent, because they consider themselves morally obliged to do so.[63] The process of legal development surrounding euthanasia in the Netherlands commenced when doctors began publicly to declare that their primary allegiance in such matters was to their professional obligations rather than to the legal prohibition. The professional requirements they worked out to govern their contra legem practice were later adopted by the courts and ultimately by the legislator, and now form the basis of Dutch euthanasia law.[64]

Pursuing Hart’s (1961) idea, it would be interesting to know more about the secondary rules that obtain on the medical shop floor: rules that identify the primary rules that bind, rules of change, and rules governing the selection, interpretation and application of rules.[65] On the one hand, it seems pretty clear that there is nothing like [page 32] a full-fledged ‘legal’ system to be found there, with institutionalized ways of performing these functions. On the other hand, it is equally clear that there is far more of this kind of thing on the medical shop floor than, for example, in the economic-sector SASFs described by Moore (1973) and Macaulay (1963). From the highest (national) to the lowest (ward) level, doctors are familiar, for example, with the explicit creation of new rules. At the ward level, to use an example from our own research, the small team that works in an intensive care unit may agree that whenever abstention from further life-prolonging treatment is being considered, one member of the team will be appointed to be responsible for all communication with the patient’s family. At the national level, the professional associations of psychiatrists, neonatologists, nursing-home doctors, and others, have issued reports on the criteria and procedures applicable to various sorts of medical behavior that shortens life. Local protocols on the subject are a common feature of Dutch medical institutions. In short, medical SASFs exhibit a range of differentiated ‘legislative’ behavior.

The mobilization of externally-oriented secondary rules is a common low-visibility feature of everyday life. What is really going on may be concealed by the mundane expressions, which accompany it. Ellickson, for example, reports a ‘choice-of-rule’ rule among rural neighbors in Shasta County, California, as follows: “Being good neighbors means no lawsuits.” Ellickson interprets this as a rule requiring the “informal resolution of internal disputes” (1991: 251). Many of the common moves in a moral discussion are similarly implicit appeals to various externally-oriented secondary rules that determine when ‘law’ is or is not relevant to a particular issue.

Professional groups, like many other SASFs, have strong rules governing the circumstances under which mobilization of external rules is (in)appropriate even if professional rules are neutral or even negative with respect to the behavior concerned. This explains why medical mistakes, for example, are rarely reported by medical professionals, despite various sorts of formal obligation to do so.[66] Doctors [page 33] who personally regard euthanasia as impermissible and properly subject to criminal prohibition may nevertheless consider it professionally inappropriate to report a colleague. In fact, very few cases of euthanasia or related behavior have ever been reported by a medical professional other than the doctor directly concerned.

Professional ‘sovereignty’ rules may ‑ in the view of the professionals concerned ‑ allocate legitimate power to regulate particular matters exclusively to the profession itself. The reaction of doctors to the efforts of the Dutch state to regulate the practice of euthanasia is heavily influenced by the historical conviction of the Dutch medical profession that it is professional ethics, not law, that govern the relationship between doctor and patient.[67] Until very recently, law was simply not [page 34] seen as competent with regard to medical decision-making in connection with the end of life.[68] In the proceedings surrounding the bill to legalize euthanasia recently enacted by the Belgian parliament, a representative of the medical profession expressed opposition to legalization because it would entail legal regulation of a practice that Belgian doctors engaged in without interference from the legal authorities and therefore subject only to professional ethics.[69]

The choice-of-rule rule of a SASF may point to an external rule as applicable in the circumstances,. In such a case, enforcement of the external rule can take place through the normal enforcement mechanisms of the SASF itself.

Until recently, the role of SASFs with regard to the effectiveness of legal rules was mostly seen in negative terms: self-regulation by SASFs was considered an obstacle that legislation must overcome to be effective. This has been the message of a great deal of research on the effectiveness of legislation, both by anthropologists in rural, Third World settings and by sociologists in western settings.[70] But as Moore [page 35] observes, it would be wrong to assume that the relationship between state law and local self-regulation is always or simply one of resistance by the latter to the effectiveness of the former, for

the various processes that make internally generated rules effective are often also the immediate forces that dictate the mode of compliance or noncompliance to state-made legal rules (Moore 1973: 721).

The assumption that legislation would generally be effective if it were not for the resistance it encounters from SASFs wrongly attributes an innate social effectiveness to legislation. It overlooks the fact to which Moore calls attention, that in many cases legislation is ineffective not because of local resistance but simply from lack of active local support.[71]

There are many striking examples of external law being enforced by the social control of SASFs. The extraordinary effectiveness in recent years of various anti-smoking measures, despite an almost total absence of legal enforcement, is a good example of the phenomenon.[72] The job of enforcement was assumed by local social control, which has proved highly effective. Such local-level control evidences itself principally, as in the anecdote with which this article began, in low-visibility, largely implicit mobilization of a sort that has not attracted much attention from sociologists [page 36] of law, such as: ‘would you mind moving to the smoking section’ or ‘shall we have a smoking pause in the middle of the meeting?’

When can we expect the externally-oriented secondary rules of an SASF to lend the support of local social control to the enforcement of external law? At least one important part of the answer is suggested by the example of anti-smoking measures. The SASFs, which legitimately[73] concern themselves with such matters, such as the general community including both smokers and non-smokers, lack much by way of internal secondary rules constitutive of a differentiated legislative function; they therefore have a hard time adjusting their behavioral rules to changing circumstances. Their members may therefore for a long time be frustrated and confused with respect to the behavioral demands that they can legitimately impose on each other. Every non-smoker has experienced this situation in years past: in meetings, social gatherings, elevators, etc. In such a case, a SASF like the state, which does have the capacity for differentiated legislation, can, by adopting a clear rule on the matter, clarify the normative situation. The external rule then has a good chance of being adopted for internal purposes and hence of effectively regulating behavior even in the absence of external enforcement.[74]

Of course, it is essential in such cases that the external rule be adapted to the internal normative situation of the SASFs concerned. By contrast with smoking restrictions - supported, argue Kagan and Skolnick (1993), by the general social rule, widely accepted in the United States, against exposing others to health risks ‑ restrictions on alcohol and other drugs seem rarely to succeed in securing the active support of informal social control. Similarly, from everything we know about the social fields that comprise the Dutch health-care system, it seems clear that a simple prohibition of euthanasia would be rejected by the externally-oriented secondary rules of the medical profession and therefore would enjoy no support from local social control. Without such support, an external legal prohibition may be an attractive symbol,[75] but it will not have much effect on behavior.[76]

The stance a SASF takes toward an external rule is by no means a static given. The content of the group’s externally-oriented secondary rules may be subject to the internal political dynamics of the group. As Macaulay (1963) emphasizes in the context of the contracting behavior of manufacturing firms, a SASF is usually not internally homogeneous in its dealings with the world around it: different sub-groups pursue different interests and some of them may have ‘internal’ reasons (such as their power relationship with other sub-groups) to support the use of ‘external’ rules and institutions. Collier’s (1976) study of legal change in Zinacantan, a Maya Indian village in Mexico, illustrates in detail the dynamic nature of the relationship between external legal rules and the internal politics of a SASF: resistance to and embracing of state law are alternating strategies in the internal struggles for power between different generations of political leaders in a local community.

In short, the effectiveness of external rules in accomplishing rule-following behavior on the shop floor often requires local support. Local support may or may not be forthcoming. Securing it may require settling for something less than or different from what the external legislator would have preferred. The ‘ideal rule’ the legislator would like to see implemented is not always the rule that in practice will produce results. The choice between political correctness and effectiveness may sometimes be a painful one, especially when the ideal rule incorporates important ideals like equality, respect for life, and the rule of law. But the alternative to settling for less than one would have liked may often be that one accomplishes less than would have been possible.

4.4. The Social Organization of the Shop Floor and the Social Working of Legislation

Not all motives for following or not following a legal rule originate from the shop floor. Drunk driving is undoubtedly deterred to some extent by the risk of injuring oneself and one’s passengers, the risk of being caught by the police, the risk of civil liability, etc. Similarly, quite apart from shop-floor rules of the sort we have been considering, the time and hassle involved in reporting a case of euthanasia undoubtedly deters some doctors from performing it or, when they do do so, from reporting it as such. A rather different example: It appeared to our field researcher that a highly personal reverence for nature induces some farmers in the Ecuadorean Amazon not to cut down trees where others would have done so. And so forth. But merely individual characteristics and personal motives are not of great importance for the overall success of a legal rule in producing rule-following behavior. Furthermore, even an ‘individual’ factor such as sensitivity to the risk of sanctions is heavily influenced by whether local SASFs attach social stigma to a person on whom official sanctions have been applied; and a similar social element will usually be present in other ‘personal’ reasons for following a rule. In short, most factors relevant to the social working of law originate from or are shaped by the shop floor.

We have been concerned in this section with the various sorts of local rules and how they bear on whether, when and how external legal rules are followed on the shop floor. In the following section we will consider the process of mobilization of law and examine how SASFs and their rules dominate every aspect of the process. If the argument proves successful, the claim that a theory of legislation must be a sociological theory will have been borne out.

So far, we have considered (in section 3) the idea of social working as an alternative to instrumentalism and have looked (in section 4) in a general way at the place where social working takes place: the shop floor. The social organization of the shop floor we have described in terms of semi-autonomous social fields. We have paid special attention to the complexity of the regulatory situation on the shop floor and in particular to the internally- and externally-oriented secondary rules of SASFs that regulate how actors do things with the many primary rules available to them. It is time now to look more specifically at how and when, in such a context, a legal rule becomes something more than black marks on paper because somebody uses it.

‘Mobilization’ refers in the most general sense to doing something with a rule, that [page 39] is, to rule following in one of the variety of ways this can be done. In the context of the theory of social working, mobilization is thus a far more encompassing concept than in the idea of ‘mobilization of law’, where it refers to the process by which a legal institution becomes seized of a case.[77] Mobilization in that limited sense - referring to one sort of rule and one sort of use - is a special case of mobilization in the general sense intended here.

The example of euthanasia can serve to illustrate some of the variety of ways a law can be mobilized. Patients mobilize the legal rules pertaining to euthanasia when they ask their doctor whether in principle he would be willing to perform it, and more acutely when they specifically request him to do so. Doctors do so when they apply the legal rules, as they know them, to their own behavior, whether in acceding to or in refusing a request, or in choosing one or another of the less problematic alternatives. The applicable legal rules can also be mobilized by all sorts of surrounding actors (other doctors, other patients, nurses, family members, etc.). The prosecutorial authorities and the Medical Inspectorate mobilize the rules in reacting to reports (usually made by the doctor concerned) of cases of euthanasia (as well as ancillary matters such as the refusal by a doctor who is unwilling to perform euthanasia to cooperate in referring a patient to a doctor who is willing). But the rules may also be mobilized in connection with things other than individual cases of medical behavior, for example in connection with the local policy of a hospital. In that case, they may be mobilized in connection with the drafting of local euthanasia rules, in connection with related matters such as protocols concerning palliative care and life-prolonging treatment of the terminally-ill, in connection with employment contracts or liability insurance, and so forth.

All of this applies, mutatis mutandis, to mobilization of the various non-‘legal’ primary rules available for use on the shop floor - professional rules, for example. And it applies as well to mobilization of secondary rules that regulate the mobilization of primary rules.

The direct effects of a rule consist of all of the mobilization that the rule gives rise to, both when people follow it in their own behavior and when they apply it to the behavior of others. Mobilization, since it is in effect just another way of referring to direct effects, can like them be primary or secondary. In the anecdote recounted at the beginning of this article, Klijn’s fellow-passengers engage in primary [page 40] mobilization of the rule against talking in a work-compartment when, seeing the relevant icon, they refrain from talking. When Klijn has to call a fellow-passenger’s attention to the icon in order to secure compliance, or even to (threaten to) invoke the authority of a conductor, he engages in secondary mobilization. So far, everything is familiar.

Secondary mobilization can, however, be far more complex than the two enforcement options available to Klijn in the anecdote. It can take place on the spot, as when in Klijn’s anecdote he calls the attention of his fellow-passengers to the icon. But secondary mobilization can also take place far removed from the concrete situation in which behavior in conformity with the rule is at issue. We can call this common and very important possibility complex secondary mobilization. The distinction between simple and complex mobilization is undoubtedly too crude to do justice to the variety of secondary uses of rules. The reason for making it here is to draw attention to the variety of mobilization processes and the variety of places mobilization can take place - something to which one must be sensitive in studying the social effects of rules.

Complex secondary mobilization refers to the behavior of an actor who mobilizes an (external) rule as a reason for intervention in the arrangements for rule-following obtaining on the shop floor. Such intervention might take the form of creating a local rule that implements a ‘legal’ one (perhaps adapting the latter to local conditions); or it might take the form of specifying some more general social rule for the circumstances of the shop floor;[78] or it might take the form of changes in operating procedures such that conformity with the external rule will take place without local actors having to follow it at all. For example, a university administration may implement a law that prohibits discrimination in hiring by further specifying the idea of forbidden discrimination for the academic context and establishing an appointments process that includes both some local affirmative-action rules (for example, requiring qualified applicants from underrepresented groups to be invited to an interview) and some features that implement the law without requiring rule following (for example, alterations in the standard text for personnel advertisements). If their behavior is effectively structured by such a local appointments procedure, those responsible for making university appointments will conform to the requirements of the external legal rule without themselves following it, and to some extent without following any rules at all.

Simple mobilization - on the spot, on the shop floor - is fundamental. Complex mobilization, after all, itself takes place on a shop floor, albeit a different one: the shop floor where a secondary actor mobilizes both the primary behavioral rule and various secondary rules that govern its use. The mobilization there is primary and simple secondary mobilization. In section 5.2 I address systematic attention to simple mobilization. In section 5.3 I make some much more tentative observations about the very under-studied situation of complex mobilization.

The process of mobilization generally begins (and generally ends) in an arena - the shop floor - whose social organization, as we have seen, is constituted by SASFs. Because of their direct control over interaction and relationships on the shop floor, they are primarily responsible for regulating the process of mobilization, thereby largely determining the occasions on which legal rules get mobilized and the consequences that follow from mobilization.

The person who mobilizes a rule may - as we know from Klijn’s anecdote - be one of the actors whose behavior is in question, or a fellow member of a relevant SASF, or an internal or external social control specialist whose job it is to secure conformity with the rule concerned. Most mobilization of most rules takes place in the course of everyday (inter)action when the primary addressees of the rule apply it to their own behavior, or use it in explaining or justifying their behavior, or when others invoke the rule to secure conforming behavior or to justify their negative reaction to departures from the rule. Mobilization of this sort is neither articulated as such nor clearly differentiated from other aspects of the (inter)action concerned. It is therefore of fairly low visibility; empirical description is rare and theoretical analysis rather primitive.[79]

When social control specialists - in the case of legal rules: representatives of the state - are involved in the secondary mobilization of a legal rule, a distinction can be made as we have seen in section 2, between ‘proactive’ and ‘reactive’ mobilization, depending on whether the initiative to specialist action comes from the specialists themselves or from some non-specialist.

Secondary mobilization of a legal rule at the initiative of social control specialists ‑ ‘proactive’ secondary mobilization ‑ is relatively rare. There are a number of reasons for this. In the first place, extensive proactive mobilization would be prohibitively expensive. In the second place, the non-autonomy of the state often sets limits on the degree of proactivity of legal control: ‘symbolic’ legislation that honors the effective power of non-state social fields by not providing for proactive control of the behavior of their members is a common result.[80] But perhaps most importantly, as Galanter (1981) has observed, the vision of comprehensive proactive mobilization ‑ of uniform and unremitting official application of law ‑ is a highly unattractive one: social life needs the flexibility and specificity of private ordering that imperfect mobilization leaves room for. ‘Privacy’ is ‑ seen from this perspective ‑ freedom from subjection to proactive social control, the freedom to break (some) rules so long as one is discrete about it. The extent to which the state – that is to say, the SASFs of the state’s control apparatus - tolerates or even actively supports a private sphere of immunity from proactive control ‑ a socially-sanctioned space for deviant behavior ‑ is a factor of considerable importance for the social working of law.[81]

In most cases it is not a specialist but a lay member of the community ‑ one of the actors concerned or a bystander ‑ who is responsible for bringing a case to the attention of a legal institution. Most secondary mobilization is in this way reactive. Secondary mobilization of law in the case of medical practice, for example, is [page 43] almost entirely reactive. Medical disciplinary proceedings, civil liability and criminal prosecutions almost always result from complaints by patients and their families; proactive mobilization by an official responsible for maintenance of the rules applicable to medical practice is rare.[82]

The initiative of a layperson that gives rise to reactive secondary mobilization is itself a special case of secondary mobilization by non-specialists, most of which takes place in ‘trouble-less cases’ or in ‘trouble cases’ that get handled without the involvement of specialists. In the next section, I discuss factors that play a role in all processes of mobilization, primary as well as secondary, by non-specialists (including the mobilization by specialists – who in this situation are non-specialists - of the secondary rules of their own shop-floor SASFs). In section 5.2.3 I consider two aspects of the special situation of (usually reactive) mobilization by enforcement specialists.

Mobilization of a legal rule involves a process that Felstiner, Abel, and Sarat (1981) call ‘naming, blaming, claiming’. Such a conception of the mobilization process is, however, too narrow for our purposes, since it assumes that it is always a ‘victim’ who mobilizes law and that ‘fault’ in some other actor is the occasion for doing so. But as we have seen above, law can be mobilized by a far wider range of actors and in a wider range of situations than the ‘dispute’ paradigm within which Felstiner, Abel and Sarat’s discussion takes place presupposes.

In all its various forms, mobilization presupposes that a number of conditions are met:

The actor who mobilizes a rule must be informed about the facts relevant to the applicability of the rule. Official mobilizers are generally poorly informed about the behavior they are supposed to regulate (this being one of the main reasons, of course, that a legal system generally has to make do with reactive mobilization). But lay mobilizers also face the problem of what one might call ‘legal intelligence’. Thus equal pay legislation, for example, can remain unused because the beneficiaries do [page 44] not know they are being paid less than others for equivalent work.[83]

In general, the extent to which the facts are known to a potential mobilizer depends heavily upon the social organization of the shop floor: the more this approaches the situation of a ‘primary group’, the better the participants will be informed of one another’s behavior.[84] Factors such as the size, economic and social organization, and even the spatial relations[85] in a community can be of critical importance. Euthanasia in the Netherlands, for example, takes place in a number of social contexts that differ precisely on the point of access to legal intelligence.[86] Family doctors account for the bulk of all euthanasia, which - when a family doctor is responsible - takes place at the patient’s home. A family doctor generally operates solistically, and few people have to know precisely what he does in the course of treating a terminally ill patient. He is in a position to keep the most critical knowledge (e.g. concerning the rate at which morphine is administered) limited to very few, apparently reliable people.[87] In hospitals, on the other hand, decision-making is often a matter of teamwork and record-keeping highly-organized; knowledge of the exact circumstances of a patient’s death is generally available to a number of staff-members and others.

Such knowledge, however, can be an embarrassment when using it threatens internal solidarity. The very fact of teamwork, while increasing the availability of knowledge among members of the team, will tend to decrease their inclination to divulge what [page 45] they know to external authorities, even if they disapprove of what was done.[88] One may surmise, for example, that in Schwartz’ (1954) kvutza - whose internal legal intelligence was far greater than that of the moshav ‑ internal ‘criminal’ behavior was less frequently reported to the external, Israeli police or other authorities.

A common way of dealing with unwonted knowledge and the embarrassment it poses to internal solidarity is by making use of the distinction between actual, personal knowledge and social, public knowledge: one only (socially) ‘knows’ what one has learned of in a formal, socially-recognized way.[89] The following anecdote told me years ago by a medical specialist I happened to know illustrates the point:

The hospital where my informant worked was strictly Calvinist Christian and in the 1980s its official policy rejected any possibility of euthanasia. Nevertheless, there was one internist in the hospital who for many years in fact carried out euthanasia. Others in the hospital were generally aware of this fact and in appropriate cases referred patients to him. But officially no one ‘knew’ anything about what he did. His retirement posed considerable internal problems, since it meant the loss of an important facility, but one which no one formally ‘knew’ anything about and which therefore could not be openly discussed.

In short, what one knows is largely determined by social organization. But the rules of the shop floor also determine how much of what one knows, one ‘knows’ in the sense that the knowledge is socially relevant and can be the basis for the mobilization of a rule.

The next thing that is required if a rule is to be applied to behavior, is that the relevant actors (those whose behavior is in question and others who might be in a position to mobilize the rule) know about the existence of the rule and what it requires.

The problems of legal knowledge begin with the legislator. A legislator whose strategy depends on people following a legal rule, and who thereby presupposes that they have knowledge of it, will have to formulate his law in terms comprehensible to the actors concerned or he will have to ensure that it be translated from legislative into ordinary language.[90] And he will have to take steps to call the (translated) text to the attention of people on the shop floor. None of this usually occurs.

When they do transmit legal information more or less directly to those on the shop floor, governments seldom speak with a single voice, that of the legally ‘correct’ interpretation of the law. Legal knowledge within the state apparatus itself - especially the legal knowledge of what Lipsky (1980) calls ‘street-level bureaucrats’ - is often problematic,[91] and even when they know the relevant legal rules officials generally do not confront the public with legal rules as formulated by the legislature but with behavioral expectations formulated in their own terms. These may be fairly recognizable ‘interpretations’ of the legislative rule or they may be local creations bearing a rather distant relationship to anything that the legislature ever enacted. Furthermore, different agencies have their own constituencies and interests, and these are reflected in the different versions of the law they communicate to the outside world. In short, the messages from the state that various bureaucrats transmit to the shop floor can be inconsistent with one another and more or less transformed and encrusted versions of what a lawyer would consider to be ‘the law’. F. von Benda-Beckmann’s (1989: 135) observation with respect to the transmission of state law to the local level in Indonesia is in this respect characteristic:

Villagers are ... confronted with local versions of State law, which often have nothing to do with the original version. Their reaction to the local bureaucrats’ demands or decisions depends on how [page 47] they interpret the local and not the original version....

Moreover, direct communication of legal information from officials to actors on the shop floor is rare. The legislature sometimes more or less explicitly - as in the case of divorce - counts on specialized intermediaries such as lawyers to transmit legal information. But most transmission of legal information presumably takes place through non-specialized institutions such as the media, the educational system, social, religious, labor, professional and commercial associations, and so forth. In the case of advance directives (in which a competent patient specifies what sort of medical treatment he wants should he become incompetent) our research suggests that transmission of the relevant legal information by legal specialists is rare. In the Netherlands, the media, doctors and other medical actors, and the Voluntary Euthanasia Association (which distributes a widely used preprinted form) seem to be the most important disseminators of information concerning advance directives. In other countries, too, ‘right to die’ societies are often active in this area. In the United States at least some Health Maintenance Organizations make forms available to their members and hospitals are legally required to inform their patients at intake, and in Spain the Bishops’ Conference distributes a form via its Internet home page.[92]

The accuracy of what intermediaries transmit varies greatly. Intermediaries themselves have limited legal capacities and resources so that the legal knowledge they possess is likely to be imperfect. Furthermore, they frequently have their own axes to grind (in the case of the Spanish bishops, for example, advance directives are seen as an alternative to euthanasia, to which the bishops are adamantly opposed). The ‘law’ such intermediaries inform local actors about may be rather different from what they themselves ‘know’, since they tend to pass along what they consider it useful for their particular public to know. And professional intermediaries such as lawyers tend to pass along legal information in terms that fit the realities of their practice situation and their own ‘definition of the situation’.[93]

SASFs often more or less totally control the flow of legal information in their area of interest to their members. The (self-appointed) representative of a local SASF may be able to enhance his political position by subjecting the legal information to which he has privileged access to selection and manipulation before passing it along. The result is the situation of ‘systematic misunderstanding’ between representatives of the legal system and the members of local SASFs, a phenomenon well known to anthropologists of law.[94]

And even if accurate information is available to actors on the shop floor, there remains the question whether they are receptive to it.[95] Receptiveness is likely to be low when external regulatory requirements differ markedly from the normative expectations that prevail on the shop floor.[96] In short, getting information to the relevant recipient in easily digestible form does not ensure that he will absorb it sufficiently even to consider making use of it.

The result of the processes by which legal information is transmitted and received is that the legal information available on the shop floor is often sparse, vague, and inaccurate. For reasons that have been extensively discussed in the literature, organizations generally have lower information costs and access to better sources of information, and are better equipped to process and use legal information than are [page 49] individuals.[97] It is thus relatively easy for a lawmaker to communicate with organizations. For this reason, enlisting the help of organizations in the dissemination of legal information can be a more effective strategy than trying to deal directly with individuals. It does, however, have its costs: when one solicits cooperation one has to pay attention to the wishes of the other party, a fact that often gives large organizations some say on the content of the legal rules they transmit to their members. An even more ambitious (but in the same way still more costly) strategy is one aimed at complex mobilization, in which local SASFs are stimulated to adopt local regulations that implement legal requirements.

The regulation of euthanasia and other medical behavior that shortens life illustrates the problems of legal knowledge that we have identified. The level of knowledge among Dutch doctors of even the most basic rules concerning euthanasia remained low for many years. Hilhorst concluded from his pioneering research into euthanasia practice that in the early 1980s ‘euthanasia’, as a concept, played practically no role in doctors’ ‘definition of the situation’: “the word euthanasia was and is taboo in hospitals”. The relevant legal rules were hardly known or applied.[99]

The mid-1980s saw the acceptance of the legitimacy of euthanasia by the Medical Association and, at about the same time, by the Supreme Court, the prosecutorial authorities and, implicitly, by the Government and Parliament. From that point on, the legal information that needed to be conveyed to doctors became steadily more technical and detailed, specifying the narrow conditions under which euthanasia is legal and the ‘requirements of careful practice’ that a doctor who carries it out must follow. The Medical Association (and in particular its weekly journal Medisch Contact), as well as other medical and nursing associations, began to generate a stream of articles, reports, guidelines, and the like. The coverage in the general [page 50] press has also been extensive and, on the whole, fairly accurate. Nevertheless, one researcher reports that doctors are quite unreceptive to the legal information available to them: they rarely read it and even today remain remarkably poorly informed.[100]

Doctors who work in institutions do not have to get the relevant legal information directly since most hospitals and many nursing homes have internal policies on the matter.[101] The capacity of medical institutions to receive and transmit legal information in this way has been striking. By 1994 - 10 years after the Supreme Court recognized the defense of justification for a doctor charged with the crime of euthanasia - three-quarters of Dutch hospitals had a permissive institutional policy, usually in writing. In general, the national legal rules, as these have emerged over the past few years, were clearly reflected in the euthanasia policies of permissive institutions. In fact, institutional policies were frequently adopted under pressure from or in cooperation with the local prosecutor or the Medical Inspector. But the rules laid down by hospital management are to some extent ‘external’ at the level of a ward. Thus the fact that an institution has an explicit policy is no guarantee that the people working in the institution know about its existence or contents or regard it as legitimate and relevant.

To make a long story short, relative to other areas of legal regulation, the conditions in the case of euthanasia seem pretty favorable for the transmission of legal information. And in fact there have been significant improvements with respect to legal knowledge concerning euthanasia. Two national studies and other indications indicate that at least as far as the essential requirements are concerned - an explicit request, unbearable suffering, consultation with a colleague, reporting - the basic [page 51] rules governing euthanasia are widely known and enjoy general support among doctors.[102] But as we will see in the next section, whether doctors interpret the rules in the same way lawyers and policy-makers do is another matter.

Literal knowledge of a rule, in the sense that one is able to reproduce it in a reasonably adequate way, cannot produce the intended direct effect unless the terms of the rule mean the same thing to people on the shop floor that they do to the legislator. As we have just seen, the basic rules relating to euthanasia are fairly well known to Dutch doctors. But how do they interpret them?

Current legal regulation of euthanasia in the Netherlands makes use of legal concepts that seem to have a significantly different meaning in the ‘semiotic group’[104] of doctors from the meaning they have among lawyers and others who concern themselves professionally with legal control. The context in which doctors learn the meaning of concepts and apply those concepts to their own behavior and that of others, is that of professional socialization and of daily professional interaction. If key terms used in legal rules mean something quite different in that shop-floor context from what they mean to lawyers and policy-makers, then the legal messages to which doctors are exposed will mean something quite different to them too. Doctors will understand the legal rules they know in terms of the concepts familiar to them, not in the terms familiar to rule-makers.

Doctors seem in fact to think about their behavior in terms of a conceptual apparatus that is radically incongruent with that of legal discourse. Let us consider the concept of ‘intent’ as exemplary of the problem.[105] The doctor’s intent plays a key role in the [page 52] legal regulation of medical behavior that shortens life, distinguishing between legally unproblematic forms (abstention, pain relief) and illegal or heavily regulated forms (euthanasia, assistance with suicide, termination of life without an explicit request, murder).

As usually employed in legal rules, intent is an objective concept: a person is taken to ‘intend’ the natural and probable consequences of his act. In this normal legal sense, a doctor intends the death of a patient every bit as much when it results from pain relief as when it results from the administration of a euthanaticum. To escape from that conclusion, the ‘doctrine of double effect’ has been borrowed for the purposes of medical law from Catholic moral theology: if the doctor’s ‘intent’ with his intervention is to relieve pain, even if he knows to a virtual certainty that this entails shortening the patient’s life, then the death is attributable to the patient’s underlying condition and the case can be regarded as one of pain relief and not of taking life. Legal regulation in both European and common-law countries has over the past few decades come to accept this alien notion of intent.[106]

Despite this concession to medical practice, doctors’ notions about what it is they ‘intend’ to do remain radically at odds with those of lawyers, as the following case illustrates.[107] It was presented for consideration to a working group on medical ethics in the neo-natal intensive care unit of a Dutch hospital, to whose meeting I was invited. I have reason to believe that discussions very much like it could be listened to in many other Dutch hospitals.

A baby was born with severe spina bifida. According to the professional norms applicable in the Netherlands to this sort of situation, any of the various surgical measures that might have been taken would have been ‘medically futile’, so the decision was [page 53] taken to abstain from any life-prolonging treatment. The parents agreed with this decision, but were insistent that the baby not suffer any pain as a result of it.

The decision to abstain necessarily entailed the death of the baby, but it might take some weeks before this would occur. In the meantime, the baby would experience pain from the naked spinal cord and also from the hydrocephalus that normally accompanies spina bifida. It was therefore considered more humane that the baby die quickly. A heavy dose of a drug used for pain relief was administered and the baby died shortly thereafter. The death was reported as a ‘natural death’.

In the discussions in the medical ethics working group, the responsible doctors consistently described what they had done as “pain relief”. There was some discussion about changing opinions on the question whether newborn babies experience much pain - apparently in the past it was believed that this was not the case and nothing much was done by way of pain relief in cases such as this (before the advent of neo-natal intensive care, the baby always died quite quickly). Later on, opinion had changed and spina bifida was taken to involve acute pain, which was aggressively treated. But, said the doctors, they were having increasing doubts about the actual level of pain experienced by such babies. When someone asked what sort of pain relief would have been necessary just to deal with the baby’s pain, the answer was that paracetamol [Tylenol] would probably have sufficed.

This answer was completely unexpected. It led immediately to the question: “I thought you said you administered [whatever the drug was] to relieve the baby’s pain.” To which the answer was, “Yes, we did. But we also wanted the baby to die quickly and it really isn’t possible in a modern hospital just to put the baby in a drafty window and wait for it to die, the way everyone did before there were intensive cares.” And to this the reply was, “How can you call it pain relief when you yourself say paracetamol would have been enough.” Answer: “But it was pain relief: we used [drug X], which is considered very appropriate for relieving pain, but we just gave rather more than we otherwise would have done.” And so forth.

How to analyze this discussion? What were the doctors doing and [page 54] why were they doing it?[108] Were these doctors simply lying when they reported the death as a ‘natural’ one, whereas they knew that what they intended was not really just pain relief? I think that would be a facile interpretation of what was going on. ‘Pain relief’ was not simply a characterization of their intent that they used to avoid having to account for what they had done; it was the characterization they themselves used in thinking and talking about what they had done. It seemed to be, for them, the natural way to look at what had happened. Were they, then, confused about what had happened? There seems to me no evidence for this. They knew exactly what they had done and why and how the baby had died.

Conclusion: The normal meaning to a doctor of ‘intent’ is incongruent with the legal meaning of this key concept involved in analyzing criminal responsibility. The current control regime for this sort of medical behavior adopts an uneasy middle position that is not understood by doctors and that they therefore cannot and do not apply properly to their own behavior and that of their colleagues.[109]

The problem of conceptual incongruence can go beyond differences in the meanings particular terms have for different semiotic groups: the legislator may conceive the very nature of the relationship to be regulated in a way that is quite foreign to the conception of the actors on the shop floor. When, for example, the Dutch legislator recently codified the legal rights of patients as these have emerged over the past few years, the choice was made to conceptualize the doctor-patient relationship as a contractual one and the enforcement of patients’ rights as a matter of remedies for breach of contract.[110] It seems doubtful, however, whether doctors and patients [page 55] experience their relationship as fundamentally contractual in nature. The way the relationship and the respective rights and duties within it looks to them probably resembles the delivery of a public service such as education - where the terms are non-negotiable and the remedies public ones ‑ much more than it does a typically contractual relationship such as that for having one’s house painted. Since virtually all elements of the medical-care ‘contract’ are fixed (by law, professional ethics, the system for financing health care, and otherwise) there is nothing to negotiate about. Nor can the actors on the shop floor recognize in their relationship the other key elements of a contractual relationship (offer and acceptance; an identifiable contractual partner on the medical side). When the legislator gets the very nature of the relationship sociologically ‘wrong’ in this way, there seems little reason to expect that actors on the shop floor will be receptive to the legislative message, or, if they receive and in some sense ‘understand’ it, will see its relevance to their situation.

An actor who knows the relevant facts and legal rule, and (in light of the local interpretation and choice-of-rule rules) understands the rule more or less ‘correctly’ and regards it as applicable, must categorize the facts as falling under the rule in question. In the anecdote with which I began this article, the concept of a ‘work-compartment’ was apparently unclear to some of Klijn’s fellow-passengers, who claimed that holding a meeting in the compartment was ‘work’. Another colleague reports a similar experience with a fellow-passenger who conducted extensive conversations on his cellular telephone in a ‘work-compartment’ and who, when she remonstrated with him, insisted that he was ‘working’.[111]

Critical elements of current legal regulation of euthanasia in the Netherlands, for example, leave considerable room for divergent categorization of the relevant facts: the patient’s ‘voluntary, informed request’ distinguishes homicide from euthanasia; the ‘requirements of careful practice’ require ‘consulting’ another doctor; the ‘cause’ (‘natural’ or otherwise) of the patient’s death determines whether a certificate of natural death may be filed; and so forth.

The process of ‘arriving at a settled interpretation of an experience’ has been systematically analyzed by Verkruisen (1993) in connection with complaints about medical treatment. He shows that there is nothing intrinsic to an unpleasant [page 56] experience that determines whether or not the patient considers it an occasion for dissatisfaction with the doctor concerned. The patient’s ‘settled interpretation’ of the experience is the outcome of processes of social interaction in which tentative personal interpretations are checked with people in his or her social surroundings. In other words, the categorization on which one ultimately settles is largely determined by the social interpretation-criteria provided by local SASFs. One of E. Abel’s (1981) findings illustrates the point: she shows that even women who know all the facts about the unfair way they were treated in academic promotion-procedures may find it difficult to question the meritocratic pretensions of the academic community. Those with whom the victim is in contact - even the victim’s own lawyer - are likely to prefer a categorization that leaves a university’s pretensions intact.[112] The victim often settles on a ‘definition of the situation’ as one due to her own failings. This, of course, is exactly what an academic SASF expects of her.

The actor who knows the facts, is informed about a legal rule and considers it applicable, and categorizes the facts he or she knows about so that they fall under that rule, must decide to do something about the situation: adjust his or her behavior to the rule or make clear to another that this is the proper course, react in one way or another to a situation of violation, engage in preventive activities such as local rule-making, etc.

Primary mobilization at the very least involves me doing something that I might otherwise not have chosen to do - for example, paying my income tax - because a rule requires it. Alternatively, if I do not want to pay taxes, I can refrain from taxable activity or arrange my affairs in such a way that they do not attract taxation. Or I can take steps not to get caught and simply not pay.

Secondary mobilization also involves the expenditure of energy, and often personal costs or some social risk. An actor must have a sufficient motive for undertaking it and this must not be outweighed by reasons for not doing so. Aubert (1966) argues, for example, that a Norwegian housemaid had little reason to mobilize the law protecting household personnel, since it was easier to get another job, and strong reasons for not doing so, since this would have made continuation of the relationship with her mistress problematic. Macaulay’s (1963) explanation for the fact that businessmen tend not to mobilize contract law is similar, as is Ellickson’s (1991) explanation for the fact that long-term rural neighbors tend not to mobilize legal [page 57] rules in their relationships with one another.[113]

There are usually many alternatives to secondary mobilization: ‘lumping it’, avoidance, exit, mobilizing some other (perhaps ‘non-legal’) rule, jotting the incident down in one’s mental moral balance-sheet of the relationship,[114] and so forth. And if an actor decides to mobilize a given rule, there are many ways from which he can choose, varying from directly confronting another actor, through approaching some more or less indigenous authority, to getting the matter onto the agenda of an external social control agency (hospital administration, public health inspectorate, prosecutor).

Often, of course, the relevant considerations are banal: my inclination to request a fellow-passenger in the ‘no-smoking’ compartment’ of a train not to smoke is influenced by a quick assessment of how unpleasant the exchange is likely to be (am I dealing with a group of drunken football supporters or someone whose looks suggest adherence to middle-class rules of behavior?) and whether it is likely to produce the desired results (is a conductor nearby and what is the likelihood that he will support my invocation of the no-smoking rule?). In short, just because there is a will, there need not always be a way. Mobilization depends on an actor being in a position to undertake it. No one, for example, can be expected to report violations of forest-protection law to the police if the latter are - as in the case of the Amazon region of Ecuador - located hours away on an unpaved road and the offender is a member of the local community with whom one must manage to get along on a day-to-day basis. For many sorts of law the availability of (effective) legal assistance may be essential to mobilization, but the distribution of legal services is everywhere highly unequal, with the rich, the powerful and the well-organized on the whole enjoying readier access and higher-quality services.[115]

Individual characteristics may also play a role. Some people are, for example, more laconic than others about ‘rules’ and ‘rights’. Some are more assertive, have more [page 58] ‘bureaucratic competence’, than others. Of all such characteristics, an actor’s organizational situation is probably the most important. With the famous term ‘one-shotter’, Galanter (1974) identified a collection of factors that, together, impose formidable limitations on the mobilization capacity of an individual acting alone. Black (1976; 1980) shows that the relative position of an actor on the ‘organizational dimension of social space’ highly determines how much attention from legal officials his problems will attract. In this way, too, it is social organization that determines the social working of law.

Particularly interesting, from the point of view of social working, is the fact that there are social rules concerning mobilization. What is often treated as a matter of cultural or even individual attitudes (referred to with terms like ‘rights consciousness’, ‘litigiousness’, and so forth[116]) in fact often consists of secondary rules concerning the way one should deal with rule-violation. Ellickson’s ‘Being good neighbors means no lawsuits’ is a good example of this. And Verkruisen (1993) shows how the decision of a patient who thinks he has been badly treated by his doctor about whether to confront the doctor, is influenced by the reactions of those in his immediate social surroundings concerning the occasions under which mobilization is or is not appropriate.

SASFs impose sanctions on those who break the secondary rules about how and when to mobilize rules. ‘Whistle-blowers’, for example, who, in violation of the rules that obtain on their shop floor, mobilize external law, can find themselves ostracized. The following recent example (this time from Norway rather than the Netherlands) will make the point clear.

In the summer of 1999 a doctor filed a complaint with the manager of his hospital alleging that a colleague had performed euthanasia (which is strictly forbidden in Norway) by using a very high dose of Dormicum (a potent pain-killer). The patient was in the terminal phase of his illness and in great pain, and had several times earlier indicated a wish to die. The Dormicum was administered by nursing personnel pursuant to telephone instructions from Rome, where the doctor concerned was on vacation.

The manager, after consulting the Complaint Committee of the hospital, came to the conclusion that there was no question of euthanasia. What the doctor had done was pain relief. The complaining doctor thereupon filed an official complaint with the [page 59] Medical Inspectorate [Helsetilsyn]. When interviewed by the press, the Inspectorate treated the case as a tempest in a teapot.

Probably nothing further would have happened with the complaint if Norway’s most prestigious newspaper (Aftenposten) had not reported on the case, together with the allegation that such euthanasia practices regularly occurred in the hospital concerned, under responsibility of the same doctor. In the meantime, the complaining doctor’s behavior had led his colleagues in the hospital to ostracize him and he was temporarily relieved of his duties.

The Inspectorate submitted the case to an ad hoc committee of specialists, whose judgment was that there was no question of euthanasia but rather of high doses of Dormicum required for pain relief. But the composition of the committee came under attack (in Aftenposten and elsewhere): it was alleged to have consisted of friends of the accused doctor.

The complaining doctor reported the case to the prosecutorial authorities, who decided to investigate. In the meantime, the Ethical Committee of the Norwegian Medical Association appointed a prestigious committee to prepare a report on ‘terminal sedation’.

In the spring of 2001 it became known that the prosecutorial authorities had decided not to prosecute because of insufficient evidence of euthanasia. It seems that the complaining doctor never returned to the hospital and found work elsewhere.[117]

As the example illustrates, the ultimate sanction visited on a persevering ‘whistle-blower’ takes the form of ostracism. Depending on the actor’s exit possibilities, this sanction will vary from a slight irritation to a matter of life and death. In many cases it takes the form of exclusion from an economically attractive activity, whether this be the exchange of local goods and services among peasant farmers far away in the tropical forest, mutual help and flexibility among neighboring cattle farmers, or reciprocal trust and flexibility in a modern business setting.[118] The sanction lies in [page 60] the fact, in Moore’s words (quoted above), that the participants want to “stay in the game and to profit from it”.[119]

Effective legal regulation of behavior on the shop floor involves being able to intervene in ongoing social relationships that are primarily regulated by SASFs, and to do so by means of external rules that must be mobilized by members of the very SASFs ‑ participants in the very relationships ‑ that the legislator would regulate. These SASFs generally have rules about when such mobilization of external law is appropriate. Studies of SASFs confirm Moore’s assessment that such external regulation is often frustrated by the secondary rules on the shop floor. In general, the pattern of reactive mobilization conforms to the pattern of existing relationships in the SASFs concerned: reactively mobilized external law respects the internal order of SASFs and has but a limited capacity to effect change there.[120]

Whether it be reactive or proactive, mobilization by legal officials ‑ often called ‘implementation’ or ‘enforcement’ ‑ is generally a more or less prominent part of the process by which the legislator presumes (in many cases probably incorrectly) that a legal rule will be transformed into the required behavior. Black (1976, 1980) argues that not legal rules but social relationships explain the behavior of bureaucrats. On the other hand, the idea of formal bureaucratic implementation - associated, perhaps not entirely justly, with Weber - treats it as completely determined by legal rules. The ‘social working’ approach, by contrast with such polar positions, regards bureaucratic rule following, like all other rule following, as a possible but socially problematic effect of legal (and other) rules. The degree to which and the circumstances under which legal rules are followed in bureaucratic behavior are [page 61] regarded as dependent above all on the social organization of the bureaucratic shop floor.

The instrumentalist conception of the state apparatus as an ideal bureaucracy in which legal information is perfectly transmitted and implemented according to its ‘legally correct’ interpretation is known to be quite remote from the reality of implementation practice.[121] The state apparatus for the implementation of legal rules (administrative agencies, police, courts) is socially organized in a large number of SASFs, which like all other SASFs, bring forth and enforce their own rules, in particular, rules about how to deal with legal rules. The process by which legislation is interpreted and applied is thus regulated not only, probably not mostly, and certainly not ultimately by ‘legal’ but by ‘extra-legal’ rules.

The analysis of secondary mobilization by officials involves the same basic elements that have been dealt with in section 5.2.2. Two aspects of bureaucratic implementation ‑ two subjects to which the internal regulation of the SASFs which constitute the government apparatus in particular addresses itself ‑ deserve special attention: how rules are interpreted by legal officials, and what style of implementation they choose. With respect to each of these, there is a considerable literature devoted to the factors that seem to be most important.

The interpretation of legal rules and the categorization of factual situations by officials, like that of non-official actors, is only in part the logical, cognitive activity supposed by legal theory; more importantly for our purposes, it is only to a very limited extent an activity of individuals at all. To a substantial extent it is carried out according to the secondary rules of the SASFs on the shop floor where bureaucratic work takes place - the rules that determine such things as what counts as a good argument.[122]

Hawkins (1984: ch. 5) shows, for example, how in the cultural and organizational [page 62] context of enforcement of water-pollution rules, the definition of a situation as ‘pollution’ rather than as ‘dirty water’ by water-pollution officials is not determined by the applicable legal provisions but is the outcome of a categorization process in which account is taken of such factors as the environmental impact of the discharge, the discharger’s technical and economic possibilities for compliance, the internal controls of the bureaucracy within which the official works, the likelihood of public attention to the case, etc. Black’s (1976; 1980) theory implies that this process is systematically influenced by the relative social positions of those with whom officials deal, so that the behavior of large and powerful organizations is less likely than that of an individual or a small business to be categorized as ‘pollution’. Studies of environmental enforcement bear Black out.

What neither Hawkins nor Black emphasize, however, is that the factors they mention are transformed into normative considerations - rules - within a bureaucratic office. A considerable body of research has shown that the interpretation and categorization rules of the SASFs present within bureaucratic organizations set very narrow limits within which the individual idiosyncrasies of the decision-maker or the specific facts of a given case can operate.[123]

The second variable aspect of bureaucratic mobilization of legal rules is what has been called ‘style’. Black (1976) distinguishes four styles of law: punitive, compensatory, conciliatory, and therapeutic. In the literature on bureaucratic mobilization, two styles are often emphasized: a ‘compliance’ style and a ‘sanctioning’ style.[124] These correspond more or less to Black’s conciliatory and punitive styles, and I will limit the discussion to them here.

In the compliance style, officials seek, in a process of continuing negotiation with the actors concerned (e.g. industrial polluters), to secure as much compliance as is feasible. Because the relationship is a cooperative one, the officials receive more [page 63] information about relevant events than would be the case if their approach were confrontational, and they are in a position to work together with industry to reduce the risk of pollution in the future. In the sanctioning style, by contrast, officials seek, by applying the law strictly to cases of infraction, to deter future violations. The two styles are obviously ideal types. An official’s bargaining position in seeking compliance depends in part on knowledge by a polluter that the official’s repertoire also includes sanctioning. And exclusive use of a sanctioning style, in which offenders are punished after the fact but no attention at all is paid to prevention, seems hardly conceivable except perhaps when a potential offender is not knowable in advance (as is often the case with common crimes). In any concrete case mobilization style will consist of a particular mix of compliance and sanctioning.

The factors associated in the literature with the prevalence of a particular style are things such as the characteristics of the bureaucratic organization involved (selection and rewarding of personnel, structure of internal control and responsibility[125]), the demands of adaptation to the chronic imbalance between work and resources (see Lipsky 1980 on ‘street-level bureaucrats’), and social-structural relationships between the those over whom the bureaucratic organization exercises social control (see e.g. Black on factors such as stratification and relational distance). What for the theory of social working is important is that - as in the case of interpretation and categorization - these factors are generally transformed into internal rules of appropriate behavior within a bureaucratic office: secondary rules about how to apply the primary rules for which the organization is responsible. Hence, for example, while as Hawkins (1984) shows, ecological or internal organizational factors may serve to explain the preference for the compliance style among bureaucrats responsible for enforcing water-quality legislation, the officials who respond in the observed way to these factors will more immediately be engaged in following local rules.

In short: bureaucratic following of secondary rules concerning the interpretation, application, and mobilization of rules is of central importance in the theory of the social working of law. As we have seen, bureaucratic mobilization of legal rules involves actors who are simultaneously mobilizing both primary (‘legal’) and secondary (local) rules, and the behavioral effects of the latter can be absolutely crucial to the behavioral effects of the former.

Complex mobilization comes, as we have seen, in two sorts. In the first, secondary mobilizers use external (e.g. ‘legal’) rules as reasons for creating local-level rules which local mobilizers can follow without necessarily even knowing of the existence of the external rule. In the second case, secondary mobilizers use the external rule as a reason for bringing local-level structures into existence that accomplish behavior in conformity with the external rule without any rule-following by the ultimate addressees at all. While such complex mobilization is of considerable importance in the social working of law, it has received no significant empirical or theoretical attention. It is therefore only possible to make a few preliminary remarks on the subject.

The ‘legal’ rules that govern euthanasia have been given expression in more ‘local’ and ‘professional’ rules in two principal ways.[126] In the first place, many hospitals and nursing homes have, sometimes in cooperation with or even at the insistence of local prosecutors, adopted local ‘guidelines’ or ‘protocols’ governing the performance of euthanasia. In the second place, both the Royal Dutch Medical Association and several specialist associations have adopted national guidelines in which the legal rules are reflected.[127] Since these local and professional rules are the result of mobilization of the legal rules, a doctor who carries out euthanasia can conform to the formal legal rules by following the local or professional ones. He might have only the vaguest idea of what the legal rules themselves require but nevertheless by following local or professional rules do exactly what the law expects of him.

Secondary mobilizers can often implement a rule better by ingenious environmental design than by any amount of effort addressed to rule following. It is this insight that led Karl Llewellyn to praise the cloverleaf highway interchange as the “greatest legal [page 65] invention of the Twentieth Century”.[128] Another classic example with a name that is particularly suited to my purposes is the ‘sleeping policeman’: artificial speed bumps that force automobile drivers to drive slowly. And if you want to ensure that cyclists equip their bikes, as Dutch law requires, with tires containing a reflecting strip on the sides, it would be far more effective simply to ensure that no other sorts of tires are available than to direct a legislative command to millions of bicycle owners. The amount and complexity of the rule following by everyone involved in such simple examples - except the ultimate addressees of the primary rules concerned - is impressive.

An interesting phenomenon that combines both aspects of complex mobilization is the legal form that one must complete in order to accomplish some task. Forms implicitly reflect the applicable legal rules and in carrying out a task as specified on a form, one inevitably (and perhaps without realizing it) does the things required by those rules. One of the nicest examples of this is the income tax return: by accurately answering questions that derive from but do not explicitly reproduce legal rules, one can conform to those rules without even having heard of their existence. In the Dutch situation, where one can fill out a return on one’s computer (and submit it by email or on diskette), the calculation of the amount owed or to be refunded takes place on the computer screen before one’s very eyes, a whole system of complicated rules being transformed into a simple, quantitative result without the taxpayer having done anything but supply the requested information.

One part of a recent proposal[129] for decriminalized control of euthanasia is an example of the unsung potential of red tape. The proposed form for reporting a case of euthanasia would have required confirmation that particular steps had been taken (such as discussing the case with the nursing staff); these would have had to be supported by signed statements of the persons concerned. Our assumption was that a doctor, knowing that he would later need such statements, would make sure to take the required steps so that when the time came he could, without difficulty or embarrassment, complete the form.

In short, ‘red tape’ looks to be too important to the social working of legal rules to [page 66] deserve the contemptuous neglect that has hitherto been its fate even among sociologists of law.

The social working of law depends upon the mobilization of legal rules on the shop floor of social life (including the shop floors where various secondary mobilization activities take place). Whether or not mobilization takes place and, if so, in what ways, is determined by a number of conditions that have been briefly touched on in section 5.2.2. Local SASFs and their regulation of behavior are a crucial factor with respect to all of these conditions. Throughout our discussion of mobilization we have repeatedly come to the conclusion that what people do with a legal rule largely depends upon the secondary rules that obtain on the shop floor of social life.

SASFs on the shop floor are the primary locus of moral training and orientation; they are the prime source of information about legal and other rules; they determine the nature and quality of legal intelligence; they are the source of the social secondary rules which govern rule-following (including sovereignty rules, choice-of-rule rules, interpretation and categorization rules and implementation rules); they dominate the process of arriving at a definition of the situation; they determine the costs and benefits of mobilization; they resist or support the application of external law to internal relationships. And they invest vastly more resources in all this social control work than even the most totalitarian state can usually achieve.[130]

In short, it is not the legislator who determines when a legal rule will be followed in social interaction. It is society that determines when and to what extent it is regulated by law. And it does this in a highly self-regulated way.

The theory of social working deals with rules as independent variables that, via the complex processes of rule following explored in section 5, can have behavioral effects in a variety of ways. The rules themselves have so far been treated as given; no explanation for their existence or contents has been offered. However, at various points we have noted that the source and content of a rule may affect the chance that those addressed follow it. External rules that, for example, are not accepted by the sovereignty rules of a SASF or are not identified by its choice-of-rule rules as [page 67] applicable to the behavior concerned, may well encounter local resistance or not receive the local support that is essential to their being followed. The way in which a rule comes into being can be important to its success precisely because of the consequences for the status of the rule in local SASFs. For this reason, it seems important to pay brief attention to the social genesis of legal rules, and in particular to one very understudied possibility that seems likely to have a positive effect on rule-following: self-regulation.

Self-regulation is, of course, the defining activity of every SASF. But our interest here is not with the general question, how and when a SASF brings forth a rule, nor with the question how the specific contents of such a rule can be explained. Our interest is with the continuity between rule making at the level of local SASFs and rule making at the ‘legal’ level. The assumption is that legal rules that emerge first in SASFs as local rules, will for a variety of reasons have a particularly good chance of being followed: they will be locally known and adapted to the local situation, they will not affront local sovereignty rules, they will be identified as applicable by local choice-of-rule rules, and they will enjoy local support.

There are two approaches to the sociological study of self-regulation. One, represented at its best by Ellickson, explains the existence of a given rule in a social group as the result of the welfare-maximizing choices of ‘rational actors’.[131] Ellickson states the central hypothesis of his approach as follows: “members of a close-knit group develop and maintain norms whose content serves to maximize the aggregate welfare that members obtain in their workaday affairs with one another” (1991: 167). He uses the hypothesis to give a plausible explanation for a whole variety of primary and secondary rules used in the settlement of disputes over damage due to roaming cattle in a rural community in Northern California. It is a theoretically interesting and ethnographically rich account, but it answers a different question from what I have in mind here. Ellickson’s hypothesis is functional: the existence of a rule in a group is explained by the fact that it would be good for such a group to have such a rule. What Ellickson does not do is address the question: how does a group acquire a rule that it would be good for it to have?[132]

Assuming, then, that the welfare-maximizing effects of a rule (or at least its lack of harmful effects) explain its survival, at least in the long run, how can we explain the genesis of the rule? Just as the theory of social working seeks to understand the process of rule following, so would a theory of the social genesis of rules seek to describe and explain the process of self-regulation.

The literature with respect to the legislative process is every bit as ‘top-down’ and positivist in its approach to legal change as instrumentalism is with respect to ‘effectiveness’.[133] The ‘new institutionalism’ reflected in particular in the writings of Edelman,[134] is an interesting exception. In a number of articles dealing with the response of large organizations to the legal requirement of non-discrimination in employment, Edelman shows how the substantive requirements of the law were transformed within organizations into procedural requirements, which in turn were later accepted by the courts as evidencing compliance with the law. In effect, to use Edelman’s terminology, the ‘endogenous’ rules of organizations had become law, supplanting the ‘exogenous’ legislative rules. Edelman’s cycle of legal change begins with a top-down legislative input but ends with the bottom-up development of rules that, at yet a later stage, gain judicial acceptance.

Our research on the regulation of euthanasia and other medical behavior that shortens life has confronted us with a repeated pattern in which legal change begins on the shop floor itself, without any initial legislative input, and only at a much later stage gets adopted by the courts and, ultimately, by the legislature. Briefly, the story is as follows. At the level of publicly visible legal development, Dutch law on euthanasia began to develop around the time of the Postma case in 1973. In that first prosecution of a doctor for killing a patient at the latter’s request, the trial court asked the Medical Inspector to testify concerning opinion in the medical profession with regard to the legitimacy of euthanasia. His testimony included the first explicit formulation of what later became known as the ‘rules of careful practice’ applicable to euthanasia. According to the Inspector, these were widely accepted among Dutch doctors. The court accepted them as defining the conditions under which euthanasia [page 69] could be legally justifiable.[135] The same conditions were later followed in other lower-court cases, were affirmed and elaborated in reports of the Medical Association, formed the basis for prosecutorial policy, were implicitly recognized (as requirements of medical ethics) by the Supreme Court when it held in 1984 in the Schoonheim case that euthanasia can be legally justifiable, formed the basis during the 1980s and 1990s of various unsuccessful legislative proposals, were accepted in the meantime by the Government and by Parliament as stating the conditions for legal euthanasia, and ultimately formed the basis of the formal legalization of euthanasia in 2001.[136]

Similar publicly-visible processes of legal development later took place with regard to related forms of medical life-shortening behavior such as termination of life in neonatology and assistance with suicide in the case of psychiatric patients. The process begins with reports by professional bodies formulating substantive and procedural rules; in a later stage these reports are followed in judicial decisions in criminal cases and in prosecutorial policy. The Government and Parliament tacitly accept this legal development and at some later stage the whole process receives legislative ratification.[137]

With the partial exception of euthanasia, in all of these cases the publicly visible process of legal change began with self-regulation by the medical profession at the national level. In fact, however, this national-level regulatory activity must have been preceded by normative change at lower levels within the profession, and ultimately on the shop floor itself: one can hardly imagine a committee at the national level making up new rules out of whole cloth.

Our research is currently directed to the shop-floor level of professional self-regulation of medical behavior that shortens life, in particular in neonatology and in intensive care. In this research we attempt to reconstruct a process in which normative development begins at the level of the ward in response to a whole variety [page 70] of possible stimuli: new problems for which existing rules give inadequate guidance (in the medical case due in particular to technological development); the risk of criminal prosecution or civil liability; pressure from outside (local prosecutors) or above (hospital management); moral entrepreneurship; imitation of the neighbors. Dutch normative development in the area of medical behavior that shortens life affords examples of all of these. In a later stage, so we suppose, professional associations set about achieving national consensus on the applicable rules. It is these national rules that in a still later stage are adopted by the courts as the law governing the behavior concerned.

But how does a new rule first emerge on the shop floor?[138] Existing notions about normative change tend to adopt a contractarian model: confronted by a new problem, the people concerned get together and ‘agree’ on a rule to deal with it. Dutch doctors do indeed tend to speak about their local-level rules as ‘agreements’. In some cases - as when the motive for change comes from outside the shop floor itself - this may adequately characterize what occurs. But normative change along the lines of the contractarian model is, I believe, far less common than the frequency of contractarian language suggests: if one asks a few skeptical questions the supposed ‘agreements’ turn out in fact to be rather rare. How then does normative change take place?

The idea that there may be quite another process of normative change at work - one that according to our current working hypothesis is more frequent and fundamental than processes of ‘agreement’ - arose out of discussions concerning a particular case described to us by a respondent who works in an Intensive Care Unit of a major hospital.[139]

The ICU concerned has four doctors, who work in shifts and are rarely present together. At any given time the doctor present is responsible for all treatment decisions. At the time the case arose, the principle was accepted by all of them that when it has become ‘futile’, life-prolonging treatment such as breathing and cardiac support can be stopped and the patient allowed to die.^[[140]] The [page 71] decision to do so is up to the doctor in charge. One day, the doctor in charge at the time decided to stop the treatment of a particular patient. He did so in a way and at a time, such that the death of the patient foreseeably would not occur until the next shift. This led to a conflict with the other three doctors,^[[141]] the resolution of which was an explicit agreement that while stopping ‘futile’ treatment is up to the individual doctor, he should if possible do it in such a way that the patient dies on his own shift. This local rule has since been passed along to successor doctors, none of the original parties still being present on the ICU concerned.

How should we interpret such a case? It comes about as close to a ‘legislative act’ as anything we are likely to find in everyday life. But is it really the case that those concerned created a new (welfare-maximizing) rule out of whole cloth? A bit of reflection suggests that that was not the case. The new rule emerged from a conflict, and a conflict presupposes an already existing rule that someone is regarded as having violated. The unspoken but latent rule in this case, I believe, is the very general one that if a body makes a mess, a body should clean it up. Such a rule is drilled into every normal child by its parents: ‘It’s OK if you play with the Lego on the living-room floor, but you have to put it away when you are done.’

In short, what seems to have happened in the case described was not the making of a new rule, but the mobilization of an old one. The bit of explicit ‘legislation’ that took place was not an act of normative creation but one of specification. ‘Clean up after yourself’ became specified for the situation of the ICU as ‘stop treatment in such a way that the patient dies on your shift [and you have to deal with the results yourself]’. Normative change may often be much less genuinely creative than is usually supposed. Probably it typically looks more like the change that takes place in a system of common-law precedent, whereby people by constantly using the normative material available to them, gradually, by processes of specification (and of distinguishing, recognizing exceptions, etc.), produce ‘new’ rules.[142]

This case and the speculation to which it gives rise suggest that normative change is [page 72] often (if presumably not always) a matter of mobilization of existing rules.[143] One of the effects of existing rules is that in everyday social interaction they generate ‘new’ rules. The continuity between ‘law’ and ‘other social control’ is thus not merely one of overlapping content (a political truism often emphasized in the ‘law and morals’ debate). Nor is it done full justice by the interesting idea that law and non-legal rules should be studied in the manner of comparative law, since ‘legal’ reasoning and everyday ‘moral’ reasoning are structurally very similar, not the radically different things that both lawyers and laymen are given to suppose.[144] Nor is the continuity of legal and non-legal rules limited to H.L.A. Hart’s (1961) idea that forms the basis for much of the analysis of mobilization in this article: that all formal ‘law’ is grounded in non-’legal’ social control. In addition to all of this, there is continuity in the process of normative change such that it will often be a misguided and impossible enterprise to want to say of a given change going on before one’s eyes that it is ‘legal’ or not ‘legal’. If I had not long ago been converted for other reasons to the legal pluralist point of view, the realization that not only in their social working (as I have emphasized in this article) but in their very genesis, ‘legal’ rules are so inextricably bound up with ‘non-legal’ rules, would surely have done the trick.

The theoretical distance between the social working of rules and the social genesis of rules is thus far smaller than is generally assumed. This conclusion calls into question the fundamental, if rather latent, assumption with which this article, following in this case in the footsteps of instrumentalism, began. This is that a sociological approach to legal rules must choose between regarding them as a dependent variable - seeking, that is, to explain legal change - and treating them as an independent variable, an explanation for the effects they have on behavior. It has always seemed to me a matter of basic scientific mental hygiene to keep those two questions separate: to be clearheaded about what it is one is trying to explain. If, as I now believe, rules are on both sides of the explanatory equation ‑ if new rules are an important part of the behavioral effects of existing rules - then it is obviously impossible to maintain that position in general and as a matter of principle.

To return now to the social working of the rules that emerge from a bottom-up process of self-regulation: the hypothesis is that legal rules that originate in this way in the SASFs where the behavior they regulate takes place, have a particularly good chance of being followed. Since, as I have just concluded, the emergence of rules is often itself a matter of mobilization of existing rules, the two theories - of the social [page 73] working and of the social genesis of rules - fit together very nicely.[145] Their perspective is bottom-up and the processes they seek to explain take place in the context of SASFs. In a double sense, then, it is not the state that regulates social life, but society that regulates itself.

In this article I have given no more than a preliminary overview of a sociological approach to legislation, an approach that seems more promising than instrumentalism. It is preliminary in the sense that it pretends no more than to indicate a number of elements of what must ultimately be an explanatory theory of rule following.

A theory is an answer to a question, and it is therefore of the greatest theoretical importance how that question is formulated. The instrumentalist approach to legislation, which in section 3.1 was examined and found wanting, poses the question as follows: ‘Under what conditions does a legislated rule have the intended social effects?’ In section 3 we have seen that a general theoretical answer to such a question is impossible and that the instrumentalist approach entails a number of (latent) assumptions that are sociologically untenable. The generally recognized failure of the instrumentalist project is thereby accounted for.

The question with which the social-working approach begins is: ‘Under what conditions do people follow a (legislated) rule?’ The most important claim of this article is that it is possible to give a general, theoretical, sociological answer to that question. The argument begins with the assumption that light can only be shed on the matter if we approach the way rules work in social life from the bottom up. The essential idea of the proposed approach is that rule following is itself a socially regulated phenomenon. The bulk of the article simply presents the fruits of having tried over a period of years to think about rule following systematically from such a perspective.

The idea of approaching legislation in terms of its social working brings a certain degree of order and a rather high degree of generality to a range of (social scientific) [page 74] experience with legislation. But at this stage the approach offers not much more than a conceptual paradigm. While bits and pieces of existing theory have been incorporated into the social working approach, what has been presented in this article cannot claim to be an explanatory theory that makes it possible to predict and explain the social working of legislation. Few testable general propositions have been ventured. What has been offered is only an initial survey of the factors that the approach suggests are important. The survey is intended as a point of departure for the formulation of theoretical propositions.

In the view taken in this article, a theory of the social working of legislation is part of an overall theory of the social working of legal rules. Most of the analysis of the social working of legislation is also applicable to legal rules of other provenance (case-law, customary law, etc.).[146] Mutatis mutandis, it is also applicable to non-'legal’ rules.

A theory of the social working of rules will in turn simply be part of a more comprehensive theory of social control generally. This larger theoretical context within which the discussion here is conceived has allowed me to assume that two theoretical problems will be solved elsewhere. One of these concerns the implications of the fact that the specific source of a particular rule is law, more particularly, legislation. The discussion in this article was limited to this case.

A second, more fundamental theoretical problem that I did not try to solve here concerns the idea of ‘following a rule’. Although I am convinced that this is both the most obscure and the most fundamental theoretical problem for a theory of social control (and hence of law and legislation), for present purposes I have simply assumed that the idea is conceptually coherent and can be empirically operationalized.

The reader may have begun to object, somewhere along the line, to the fact that in dealing with rule-following the concept of a SASF seems to function as a sort of cure-all, invoked in connection with practically every matter that arises. I am not inclined to be apologetic about this. It is simply the result of putting the shop floor of social life at the center of attention. The concept of a SASF is nothing more than a tool for the description and analysis of the social organization of this concrete social situation. The term itself solves no problem, theoretical or otherwise. Its frequent use reflects only a conviction concerning the place where the solutions are to be found.

An anonymous referee of an earlier version of the article was troubled with what seemed to him/her the argument’s “over-legalization of social life” - the suggestion that “all settings are governed by rules”. The short response to this is that the theory of social working is a theory about rules - in particular, about rule-following - and makes no pretense to offering an all-encompassing account of all social behavior; that while the defining characteristic of an SASF is that it produce and maintain rules, this by no means assumes that all behavior on the shop floor of social life is rule-bound behavior. This response is correct as far as it goes, but misleading. One of the implicit convictions that underlie my argument is that social behavior is, to a far greater extent than is often recognized, rule-guided behavior. Others have emphasized that social rules are, if one looks closely, ubiquitous.[147] My claim is less modest and it would be disingenuous of me to deny that the hypothesis that whether and how one particular rule is followed is to a very important degree a matter of other rules that bear on the situation, lies at the heart of my argument. In short, the criticism contains an element of truth, one calling for ethnographic research and analysis.

And finally, a parting observation on a matter that has not been my main concern: the usefulness of the sort of theory I have in mind. My own central interest has been scientific, not practical: to lay the groundwork for a genuinely sociological theory of the social effects of legislation by considering one fundamental question: How and under what circumstances does legislation influence behavior on the shop floor of social life? Contributing to the effectiveness of legislation has not been my primary concern. But there is, as Leibnitz said and Marx repeated, nothing so practical as a good theory. As the theory of social working becomes more refined and comes to include well-tested explanatory propositions, it will enable us to ‘know’ a great deal about the likelihood that a given piece of legislation - such as that concerning euthanasia - will affect the behavior of those to whom it is addressed, and to know this before the legislation is enacted and without having done expensive and time-consuming empirical research. If so inclined, legislators might use the theory in order to adapt legislation to the social contexts in which they would like to influence behavior. In this sense the theory will, by contrast with instrumentalism, be both scientifically interesting and potentially useful. One must, however, never forget Lounsbury’s admonition to “view with profound respect the infinite capacity of the [page 76] human mind to resist the introduction of useful knowledge”.[148] Without, at least as sociologists, being discouraged by it.

1981 ‘Collective protest and the meritocracy: faculty women and sex discrimination lawsuits.’ Feminist Studies 7: 505-538.

2001 ‘Euthanasia: the process of legal change in Belgium. Reflections on the parliamentary debate.’ Pp. 29-47 in A. Klijn, M. Otlowski and M. Trappenburg, ed., Regulating Physician-Negotiated Death. The Hague: Elsevier bedrijfsinformatie bv.

1993 Deciding Who Lives: Fateful Choices in the Intensive-Care Nursery. Berkeley: University of California Press.

1999 Gelijke behandeling: regels en realiteit [Equal Treatment: Rules and Reality]. The Hague: Sdu Uitgevers.

1989 ‘Scape-goat and magic charm: law in development theory and practice.’ Journal of Legal Pluralism 28: 129-148.

1988 ‘De ongrijpbaarheid van informeel recht’. Recht der Werkelijkheid 9/1: 91-99.

1973 ‘The mobilization of law.’ Journal of Legal Studies 2: 125-149; reprinted pp. 41-63 in The Manners and Customs of the Police. New York: Academic Press, 1980.

1980 ‘Dispute settlement by the police.’ Pp 109-192 in The Manners and [page 77] Customs of the Police. New York: Academic Press.

1993 ‘Social control as a dependent variable.’ Pp. 1-26 in D. Black, The Social Structure of Right and Wrong. San Diego: Academic Press.

1980 ‘Mobilisierung von Recht [Mobilization of law].’ Zeitschrift für Rechtssoziologie 1980/1: 33.

2002 Consequences. The Impact of Law and its Complexity. Toronto: University of Toronto Press.

1965 ‘The differing realms of the law.’ Pp. 33-42 in L. Nader, ed., The Ethnography of Law. American Anthropologist vol. 67, nr. 6, pt. 2.

1991 ‘A focus theory of normative conduct: a theoretical refinement and reevaluation of the role of norms in human behavior.’ Advances in Experimental Social Psychology 24: 201-234.

1976 ‘Political leadership and legal change in Zinacantan.’ Law and Society Review 11: 131-163.

1969 Discretionary Justice: A Preliminary Inquiry. Urbana: University of Illinois Press.

1978 The Social Order of a Frontier Community. Jacksonville, Illinois 1825-70. Urbana: University of Illinois Press.

n.d. ‘Constructing legalities: socio-legal fields and the endogeneity of law.’ Forthcoming in W. Powell and D. Jones, ed., Bending the Bars of the Iron Cage: Institutional Dynamics and Processes. Chicago: University of Chicago Press.

2002 ‘Legality and the endogeneity of law.’ In R. Kagan, M. Krygier and K. Winston, ed., Legality and Community: On the Intellectual Legacy of Philip Selznick. IGS Press/Roman & Littlefield.

2001 ‘Diversity rhetoric and the managerialization of law.’ American Journal of Sociology 106: 1589-1641.

1993 ‘Internal dispute resolution: the transformation of civil rights in the [page 78] workplace.’ Law and Society Review 27: 497-534.

1999 ‘The endogeneity of legal regulation: grievance procedures as rational myth.’ American Journal of Sociology 105: 406-454.

1936 The Fundamental Principles of the Sociology of Law. Cambridge MA: Harvard University Press (first German edition 1913).

1976 ‘Private ordering through negotiation: dispute-settlement and rulemaking.’ Harvard Law Review 89: 637-681.

1991 Order without Law. How Neighbors Settle Disputes. Cambridge MA: Harvard University Press.

1989 The Cement of Society: A Study of Social Order. Cambridge: Cambridge University Press.

1969 Law without Precedent. Legal Ideas in Action in the Courts of Colonial Busoga. Chicago: University of Chicago Press.

1981 ‘The emergence and transformation of disputes: Naming, blaming, claiming...’ Law and Society Review 15: 631-654.

1975 Doctoring Together: A Study of Professional Social Control. New York: Elsevier.

1964 The Morality of Law. New Haven: Yale University Press [revised edition 1969].

1969 ‘Human interaction and the law.’ American Journal of Jurisprudence 14: 1-36.

1974 ‘Why the ‘haves’ come out ahead: speculations on the limits of legal change.’ Law and Society Review 9: 95-160.

1981 ‘Justice in many rooms: courts, private ordering and indigenous law.’ Journal of Legal Pluralism 19: 1-47.

1998 ‘Studying how law matters: an introduction.’ Pp. 1-14 in B. Garth and A. Sarat, eds., How Does Law Matter?. Evanston, Ill.: Northwestern University Press.

1989 ‘‘Plunderers into Planters: Zeeland oystermen and the enclosure of the marine commons.’ Pp. 89-105 in J. Boissevain and J. Verrips, eds., [page 79] Dutch Dilemmas: Anthropologists Look at the Netherlands. Assen: Van Gorcum.

1961 Asylums: Essays on the Social Situation of Mental Patients and other Inmates. Pelican: Harmondsworth.

1982 ‘Nature is to culture as praying is to suing: legal pluralism in an American suburb.’ Journal of Legal Pluralism 20: 17-35.

1984a ‘The division of labor in social control.’ Pp. 37-70 in D. Black, ed., Toward a General Theory of Social Control (deel I, ‘Fundamentals’). Orlando, Florida: Academic Press.

1984b ‘The general theory of litigation ‑ a first step.’ Zeitschrift für Rechts-soziologie 5: 145-201.

1984c ‘Village justice in the Netherlands’ (review article on Van den Bergh 1980). Journal of Legal Pluralism 23: 17-42.

1985 ‘Four laws of interaction in circumstances of legal pluralism: first steps toward an explanatory theory.’ Pp. 217-227 in A. Allott and G. Woodman, ed., People’s Law and State Law: The Bellagio Papers. Dordrecht: Foris Publications.

1986b ‘What do Dutch lawyers actually do in divorce cases, with special reference to custody and visitation?’ Law and Society Review 20: 135-175.

1986c ‘De rechtspsychologische benadering van rechterlijk beslissingsgedrag: enkele kanttekeningen bij recent Nederlands onderzoek’ [The legal-psychological approach to judicial decisionmaking: some comments in connection with recent Dutch research]. Mens en Maatschappij 61: 199-208.

1990 ‘De sociale werking van rechtsregels en het emancipatoire potentiëel van wetgeving [The social working of legal rules and the emancipatory potential of legislation].’ Recht der Werkelijkheid 1990/2: 27-46.

1991 ‘Legal pluralism and the social working of law.’ In A. Soeteman e.a., ed., Coherence and Conflicts Concerning the Law. Zwolle: W.E.J. Tjeenk Willink.

1995a ‘Legal pluralism and the theory of legislation - with special reference to the regulation of euthanasia.’ In H. Zahle and H. Petersen, ed., Legal Polycentricity: Consequences of Pluralism in Law. Aldershot: Dartmouth Pub. Co.

1995b ‘Normative and ‘rational choice’ accounts of human social behavior.’ European Journal of Law and Economics 2: 285-299.

1996a ‘De sociale werking van recht [The social working of law].’ Pp. 469-513 [page 80] in J. Griffiths, ed., De sociale werking van recht. Nijmegen: Ars Aequi Libri.

1999a ‘Legal knowledge and the social working of law: the case of euthanasia.’ Pp. 81-108 in H. van Schooten, ed., Semiotics and Legislation. Jurisprudential, Institutional and Sociological Perspectives. Liverpool: Deborah Charles Publications.

1999b ‘The social working of anti-discrimination law.’ Pp. 313-330 in T. Loenen and P.R. Rodrigues, ed., Non-Discrimination Law: Comparative Perspectives. Dordrecht/Boston/London: Kluwer Academic Publishers.

2000 ‘Self-regulation by the Dutch medical profession of medical behavior that potentially shortens life.’ Pp. 173-190 in H. Krabbendam en H.M. ten Napel, ed., Regulating Morality. Antwerpen: Maklu

1998 Euthanasia and Law in the Netherlands. Amsterdam: Amsterdam University Press.

1986 ‘Social feuding and law in the Finnish gypsy community: blood feuding as a system of justice.’ Journal of Legal Pluralism 24: 101-125.

1963 Symbolic Crusade. Status Politics and the American Temperance Move-ment. Urbana: University of Illinois Press.

1984 Environment and Enforcement. Regulation and the Social Definition of Pollution. Oxford: Clarendon Press.

1973 ‘Trouble-cases and trouble-less cases in the study of customary law and legal reform.’ Law and Society Review 7: 585-609.

1983 ‘The concept of legitimacy in the sociology of law.’ Wisconsin Law Review 1983: 379-426.

Wereldoorlog [The Kindgom of the Netherlands during the Second World War], volumes 5, 6 and 7.

2001 ‘The legal dimensions of everyday life.’ Canadian Journal of Law and Society 16: 45-65.

1978 Regulatory Justice: Implementing a Wage-Price Freeze. New York: Russell Sage.

1993 ‘Banning smoking: compliance without enforcement.’ Pp. 69-94 in R. Rubin and S.D. Sugarman, ed., Smoking Policy: Law, Politics and Culture. Oxford: Oxford University Press.

1980 Taboe, macht en moraal in Nederland [Taboo, Power and Morality in the Netherlands]. Amsterdam: De Arbeiderspers.

1979 ‘Toward an integrated theory of imposed law.’ Pp. 289-306 in S.B. Burman and B.E. Harrell-Bond, ed., The Imposition of Law. New York: Academic Press.

2001 ‘Will doctors’ behavior be more accountable under the new Dutch regime? Reporting and consultation as forms of legal control.’ Pp. 157-178 in A. Klijn, M. Otlowski and M. Trappenburg, ed., Regulating Physician-Negotiated Death. The Hague: Elsevier bedrijfsinformatie bv.

1980 Street-Level Bureaucracy ‑ Dilemmas of the Individual in Public Services. New York: Russell Sage Foundation.

1963 ‘Non-contractual relations in business: a preliminary study.’ Law and Society Review 28: 55-67.

1979 ‘Lawyers and consumer protection laws.’ Law and Society Review 14: 115-171.

1995 ‘Crime and custom in business society.’ Journal of Law and Society 22: 248-258.

2002 Website of the Law Commission of Canada: http://www.lcc.gc.ca [page 82] (accessed 5 Aug. 2002).

1968 ‘Law as an instrument of revolutionary change in a traditional milieu: the case of Soviet Central Asia.’ Law and Society Review 2: 179-228.

1968 Law and Equal Opportunity: A Study of the Massachusetts Commission against Discrimination. Cambridge, Mass.: Harvard University Press.

1986 The Fisherman’s Problem. Ecology and Law in the California Fisheries, 1850-1980. Cambridge: Cambridge University Press.

1987 The Question of the Commons. The Culture and Ecology of Communal Resources. Tucson: University of Arizona Press.

1990 Getting Justice and Getting Even. Legal Consciousness among Working-Class Americans. Chicago: University of Chicago Press.

1973 ‘Law and social change: the semi-autonomous social field as an appropriate subject of study.’ Law and Society Review 7: 719-746.

1998 ‘Changing African land tenure: reflections on the incapacities of the state.’ European Journal of Development Research 10: 33-49.

2001 ‘The prospects of effective legal control of euthanasia in Belgium: implications of recent end-of-life studies.’ Pp. 179-194 in A. Klijn, M. Otlowski and M. Trappenburg, ed., Regulating Physician-negotiated Death. The Hague: Elsevier bedrijfsinformatie bv.

2003 ‘The non-application of the law of presumed consent to organ donation in French hospitals. A study of the failure of an attempt to modify behaviour by law’. Unpublished paper delivered to the MBPSL research group, Department of Legal Theory, University of Groningen, 11 June 2003..

1993 Voorkeursbeleid op lokaal niveau. Een rechtssociologisch onderzoek naar de sociale werking van maatregelen ter verbetering van de arbeidspositie van vrouwen bij de gemeentelijke overheid [Affirmative Action at the Local Level: A Socio-Legal Study of the Social Working of Measures to Improve the Position of Women in the Workforce of City Governments]. Groningen: Wolters-Noordhoff.

1986 De politieke strijd rondom de abortuswetgeving in Nederland [The Political Struggle over Abortion Legislation in the Netherlands]. The Hague: VUGA.

1976 Legal Choices for State Enterprises in the Third World. New York: New York University Press.

1973 Implementation. How Great Expectations in Washington are Dashed in Oakland. Berkeley: University of California Press.

1954 ‘Social factors in the development of legal control: a study case study of two Israeli settlements.’ Yale Law Journal 63: 471-491.

1998 Seeing Like a State. How Certain Schemes to Improve the Human Condition Have Failed. New Haven: Yale University Press.

1971 Internalization of Norms: A Sociological Theory of Moral Commitment. Engelwood Cliffs, NJ: Prentice-Hall.

1979 ‘The Equal Pay and Sex Discrimination Acts: their impact in the workplace.’ Feminist Review 1979/1: 37-57.

1985 ‘The concept of pela and its social significance in the community of Moluccan immigrants in the Netherlands.’ Journal of Legal Pluralism 23: 177-208.

1976 Passing On: The Social Organization of Dying. Englewood Cliffs, NJ: Prentice-Hall.

1906 Folkways. A Study of the Sociological Importance of Usages, Manners, Customs, Mores and Morals. Boston: Ginn and Company.

1995 The Role of Law in the Protection of the Tropical Forest in Ecuador’s [page 84] Amazonian Region. Groningen: Faculty of Law, University of Groningen.

1978 ‘Litigious marginals: character and disputing in a Bavarian village.’ Pp. 86-121 in L. Nader and H.F. Todd, ed., The Disputing Process ‑ Law in Ten Societies. New York: Columbia University Press.

1993 Dissatisfied Patients: Their Experiences, Interpretations and Actions. Dissertation, University of Groningen.

1994 ‘Rules for breaking formal rules: social representations and everyday rule- governed behavior.’ Journal of Psychology 128: 485-497.

2001 A Comparative Study of the Legal Status and Social Practice of Advance Directives, First Interim Report. Groningen: Faculty of Law, Department of Legal Theory.

1949 Geschiedenis van het verzet der artsen in Nederland [History of the Resistance by Dutch Doctors]. Haarlem: H.D. Tjeenk Willink & Zoon.

2001 ‘Euthanasia: the process of legal change in the Netherlands; the making of the requirements of careful practice.’ Pp. 11-27 in A. Klijn, M. Otlowski and M. Trappenburg, ed., Regulating Physician-Negotiated Death. the Hague: Elsevier bedrijfsinformatie.

1983 ‘The limits of ‘The Limits of Law’.’ Journal of Legal Pluralism 21: 129-146.

1976 ‘Responses to grievance behavior: extended cases in a fishing community.’ American Ethnologist 3: 353-373.

[1] The theory developed in this article was first adumbrated in writings that dealt primarily with other matters (Griffiths 1978, 1984a, 1984b, 1986a). I stated it explicitly for the first time (in Dutch) in connection with the question whether a significant ‘emancipatory potential’ can be attributed to legislation designed to correct social inequality (Griffiths 1990). Earlier versions of the present argument are Griffiths 1991, 1995a, 1996. So many friends and colleagues have influenced the presentation of the argument at various stages in its long gestation that it would be impossible for me to acknowledge all of them here. I am, as in the past, grateful for the critical inspiration afforded by the CHAZERAS Society. Finally, some important last-minute improvements are due to the useful suggestions of the editor of the Law and Society Review (which ultimately decided not to publish the article) and the Review’s anonymous reviewers.

[2] Witte akker, zwart zaad: Dutch metaphor for the printed page.

[3] Cf. Griffiths 1978; Bogart 2002.

[4] “The big question that confronts sociologists of law is: how do rules influence people’s behavior?” (K. von Benda-Beckmann 1988: 94) The word ‘behavior’ should perhaps be emphasized here. The suggestion that legal rules might, for example, influence how people ‘think about law’ without this having any behavioral consequences - if we suppose the idea to be intelligible at all - is simply not part of what the theory of social working of legal rules addresses itself to.

[5] The theory plays a much more explicit role in a current research program in which the focus on euthanasia in the Netherlands is expanded to include other medical behavior that shortens life, seen from an explicitly international and comparative perspective. For a description of the program and its various projects, activities and publications, see http://www.rug.nl/rechten/mbpsl.

[6] See Taale and Griffiths 1995.

[7] See Griffiths 1990, 1999b; Oden 1993.

[8] A recent evaluation of Dutch equal-opportunities legislation is explicitly based on this approach (Asscher-Vonk and Groenendijk 1999).

[9] The central place that our research concerning the regulation of euthanasia plays in the argument gives rise to a problem of sources. On the whole, I have kept references to sources in Dutch to an absolute minimum, in particular when the relevant Dutch sources are available to the interested reader in Euthanasia and Law in the Netherlands.

[10] My colleague Albert Klijn has recently applied the social-working approach to the situation of facilitative rules in his evaluative study of Dutch legislation that allows parents to choose the mother’s rather than the father’s family name for their children. See Klijn and Beijers 2002.

[11] By ‘legislated’ rules I mean rules enacted by an institution empowered to do so. It would be possible to reformulate the argument in more empirical terms (differentiation and ‘externality’ of the source of a rule) but for present purposes doing so would be unnecessarily cumbersome.

[12] A note on the use of the term ‘rule’. With the term ‘rule’ I mean a standard for behavior that is (or can be) stated as a conditional imperative and that is invoked in concrete interaction as a reason for or against particular conduct and supported by definite social sanctions (although in a given case these may be no more than general disapproval). I have avoided using the term ‘norm’ altogether - in particular, I do not follow Ellickson (1991) in using ‘norm’ to refer to non-legal rules - because it is the form and not the content or the source/status, that is important for my argument, and I want to emphasize the essential continuity, as far as rule following is concerned, between ‘law’ and the rest of social control.

[page 7]

It is of course true that a complete normative system (‘legal’ or otherwise) includes propositions other than rules, a subject to which Honoré (1987) has devoted useful attention. It is also true that rules can be contested, unequally enforced, subject to change, and the like. Taking account of all of this would have made this article even longer and more complicated without making any substantial difference to the argument.

[13] Cf. Griffiths 1978 for some tentative thoughts on that subject.

[14] Cf. Schwartz 1954; J.F. Scott 1971; Cialdini, Kallgren and Reno 1991; Griffiths 1995b for tentative contributions to a theoretical understanding of the capacity for rule following and the circumstances under which it takes place. Important to note here is the importance of the socialization process in the course of which individuals learn how to follow rules, since it underscores a point I will repeatedly emphasize in this article: how essential it is that a theory of legislation be a sociological theory.

[15] See Griffiths 1978 for these distinctions. Two other sorts of effects of rules discussed there are not central to the present argument and will not be dealt with: ‘independent’ effects, which occur whether or not there are direct effects (for instance, so-called ‘symbolic’ effects - see Gusfield 1963; Aubert 1966; Pozen 1976), and ‘unintended’ effects (which can be direct, indirect or independent).

A note on ‘symbolic effects’: It is a common response by those who strongly believe in the importance of the objectives of a piece of high-minded but apparently ineffective legislation, that despite its seeming ineffectiveness it deserves support because of its ‘symbolic’ effects. This is a curious idea, usually particularly obscure because of a failure to specify what sort of ‘symbolic’ effects one has in mind. When a rule is regarded as important despite the failure of research to demonstrate direct effects - rule-following behavior ‑ the idea of ‘symbolic’ effects might refer to hoped-for long-term effects on behavior, perhaps via an indirect route of the sort that the theory of social working calls attention to (e.g. influence on shop-floor self-regulatory activities); in that case, an appeal to the rule’s ‘symbolic’ importance is just a rather obscure way of calling attention to the fact that the jury is still out on the question of the rule’s effectiveness. On the other hand, it may be that with the word ‘symbolic’ reference is being made to effects that even in the long run are non-behavioral, a notion which, if comprehensible at all, falls outside the scope of the theory of social working of law (and, indeed, of sociology).

[16] A distinction can be made in this connection between ‘conformity’ and ‘compliance’ (or ‘obedience’). The difference is essentially a methodological one. Behavior can conform to a rule without the rule affording an explanation for it: the behavior might have occurred in the absence of the rule. In studying the effectiveness of ‘advance directives’ (and of the legislation providing for their binding force), for example, one must take account of the possibility that an advance directive may request a doctor to refrain from life-prolonging treatment that he would not have given anyway, so that the directive (and the law) cannot be regarded as responsible for the doctor’s behavior (see Vezzoni and Griffiths 2001). In the sense of the terms used in this article, mere conformity with a rule is not the same thing as following it.

[17] Years ago (Griffiths 1978) I began using the term ‘secondary’ for this sort of direct effect, simply as a way of making a desired distinction between sorts of effects. As we will see in section 4.2, this was a serendipitous choice, since secondary effects entail that a person who follows a primary rule in this way also be following ‘secondary rules’ (in Hart’s, 1961, sense): rules about rules. Various expressions used in this article (‘secondary’ effects/rules/rule-following; mobilization) thus link, analytically, the question whose rule following is at issue with the question whether secondary rules are involved.

[18] Cf. Cialdini, Kallgren, and Reno, 1991, on some of the indirect ways one person can focus the attention of another on a behavioral rule the latter ought to follow.

[19] See Black 1973 for this distinction and its importance.

[20] Schwartz (1954) gives a preliminary - but, so far as I am aware, after half a century still unique - empirical/theoretical analysis of this situation of vicarious learning to follow a rule.

[21] Cf. also Macaulay 1963.

[22] Compare on the effectiveness of anti-smoking regulations, Kagan and Skolnick 1993.

[23] F. von Benda-Beckmann 1989: 129.

[24] “The basic question guiding the instrumental approach is what social results can be attributed to law or legal reforms” (Barth and Sarat 1998: 2). Compare Bogart 2002: 114 ff. on instrumentalism and its critics. The causal conception of the connection between a legal rule and behavior is sometimes called ‘determinism’ (see Von Benda-Beckmann 1989: 130, who associates it with ‘structural-functionalism’). Von Benda-Beckmann’s criticisms of a ‘deterministic’ conception are similar to those made here of ‘instrumentalism’.

[25] Compare the situation in criminology, in which deviance was generally thought to require special explanation. Hirschi (1969) showed that it is conformity that needs to be explained.

[26] Both the Weberian tradition in sociology of law and the way in which contract law on the whole is taught (cf. Macaulay 1995), for example, assume that contract doctrine describes behavior: legally-valid contracts are entered into because being able to rely on performance is essential to a capitalist economy, and the ‘normal’ reaction to breach of contract is therefore a lawsuit to secure the value of the promised performance.

[27] See Griffiths 1978 for a discussion of the literature as of a quarter of a century ago. Cf. Woodman 1983; Von Benda-Beckmann 1989. As Woodman observed of Allott’s comprehensive survey of the effectiveness of law from the instrumentalist point of view (Allott 1980):

This book probably represents the furthest possible achievement of studies that assume that legal systems are composed primarily of imperative hypothetical-conditional norms, each created through an act of will by law-givers with a clearly defined purpose, and thereafter functioning like a more or less efficient machine to secure or fail to secure compliance with that purpose. But its inadequacies cause it to serve as a demonstration per reductionem ad absurdum of that type of study. (Woodman 1983: 140)

More recently, writers identified with the ‘new institutionalism’ in the sociology of organizations have renewed the attack on instrumentalism (thereby paying witness to the tenacity of an outmoded paradigm). Their criticism of the top-down, order-and-compliance idea of the relation of law to organizational behavior is similar to that given in this article. See e.g. Edelman n.d..

[28] The philosophical objections to such a command-conception of rules are well known (see Hart 1961).

[29] The notion of ‘the intention of the law-maker’ has also been a formidable stumbling block in the way of an empirical theory along instrumentalist lines (see e.g. Aubert 1966 and Pozen 1976 on the difference between the publicly-proclaimed and the actual intentions of legislators).

[30] This idea lies at the heart of the Liberal theory of the state and law, which has never been able to deal satisfactorily with the phenomenon of autonomous groups within a state. The solidarity of groups is reduced to voluntary, contractual relationships and the group itself to a creature of private law; alternatively, groups are seen as creatures of public law, derivative of the state and authorized and regulated by it. Since Liberal theory thus leaves no room for non-voluntary social limitations on the behavioral freedom of the individual other than those imposed by the state, the actual regulatory power of groups must be considered atavistic and deserving to be eradicated (the Jacobin and Soviet approach) or dealt with by means of fictions such as implicit consent, implicit authorization, etc. (the approach of legal theory).

[31] “Give me a place to stand and I will move the earth.”

[32] Compare J. Scott 1998.

[33] The expression is taken from Norbert Elias via Kapteyn (1980: 22).

[34] Compare Sunstein 1996. Because ‘rational choice’ approaches to human social behavior (e.g. Coleman 1990) generally neglect the essentially social nature of the phenomenon of rule following, their value for our purposes is limited (cf. Griffiths 1995b).

[35] Cf. Galanter 1981.

[36] This may be a point where the peculiarities of the American political system (and the dominant place of American scholars in the social scientific study of law) have been a hindrance to the development of a realistic theory of legislation. The everyday sense in which legislators in a European country such as the Netherlands are just ‘ordinary people’ who interact on a daily basis with other ‘ordinary people’ deprives the notion of Archimedean legislative autonomy of any empirical plausibility.

Even looked at from a legal perspective, the idea of legislative autonomy seems distant from reality. Legislation is a product of a specific legal culture and a specific legal history. It develops in a process of reciprocal influence with judicial and administrative interpretation and application of the law, supra-national legal developments, etc. The concept of the ‘law-maker’ reduces a complex historical process in which many actors and factors are involved to a simple, immediate act.

[37] A nice example of how the top-down perspective characteristic of legislators and policy-makers can lead them to conceive the social situation on the shop floor in a way that is radically wrong is given in Nowenstein’s (2003) research on the social working of the French law on organ donation. French law presumes consent to donation in the absence of explicit refusal by the donor and accords the family no power to override the patient’s (presumed) consent. However, the social reality of intensive care practice - which is where most potential donors are to be found - is one of intense involvement of the family, the patient himself being ‘brain dead’. Intensive care doctors are closest to the family of the deceased and simply unwilling to remove organs if the family objects. The French legislator’s individualistic ‘misreading’ of a social situation that is not individualistically organized shares much of the blame for the law’s ineffectiveness.

[38] I say this with the benefit of hindsight. It took Donald Black’s withering criticism of an earlier version of the argument to awake me from my theoretic slumber:

I do not believe the effects of law fall within the jurisdiction of the sociology of law....

I believe that each field of sociology should be defined by the range of variation it seeks to order, i.e., its cluster of dependent variables....

[Y]our effort to include the effects of law in the sociology of law [requires you] to take responsibility for everything on which law has an effect and for everything else that might have effects on the same elements of human behavior [such as religion, politics, recreation, etc.]. This would quickly expand your subject in a way that would seem to me unmanageable....

For these reasons, I would not venture to comment on your paper. I would regard the likely effect of anything I might say as a case of spitting in the ocean. (Letter of 14 Feb. 1996; compare Black 1993.)

As is plain from the text, I regard Black’s objection as telling only insofar as the indirect effects of law are concerned.

[39] Cf. McKay and Acheson 1987; Ellickson 1991; Van Ginkel 1989; McEvoy 1986.

[40] Cf. Griffiths, Bood, and Weyers 1998.

[41] Cf. Pozen 1976.

[42] Compare Galanter 1981 for a similar language metaphor. Where Galanter draws a lesson for the analysis of the social working of law from the relationship between a written and a spoken language, I would draw one from the relationship between a dialect and a national language. Fuller’s metaphor, thus interpreted, draws attention to the complexity and two-sidedness of the relationship.

[43] Cf. Griffiths 1986a. Ehrlich’s concept of a ‘social association’ is largely similar to Moore’s SASF, albeit less sharply focused.

[44] Compare Macaulay 1963 on the business community.

[45] Cf. Schwartz 1954.

[46] For a similar conception of the social group as the fundamental social locus of ‘law’ (in the inclusive sense for which ‘social control’ is used in this article) see Honoré 1987. For examples of the SASFs referred to in the text see: E. Abel 1981; Anspach 1993; Griffiths 1984c; Collier 1976; Edelman n.d., 2002; Ellickson 1991; Elster 1989; Freidson 1975; Goffman 1961; Greenhouse 1982; Grönfors 1986; Homans 1951; Macaulay 1963; Massel 1968; Moore 1973; Schwartz 1954; J.C. Scott 1998; Strijbosch 1985; Sudnow 1976; Sykes 1958; Todd 1978; Yngvesson 1976.

[47] Strictly speaking, the concept ‘state’ does not refer to one but to a whole complex of SASFs: legislative, administrative and judicial institutions (and their constituent units, such as a particular administrative office or a particular court) and the various socio-political fields that lie at their periphery. Reification of the ‘state’ (and ‘the legal system’ or ‘law’) ‑ common among lawyers and almost equally so among social scientists ‑ suggests a far simpler and neater state of affairs than actually obtains, and is an important obstacle to the development of theory concerning law and social organization in complex societies.

[48] In countries such as Ireland and Italy, to give but one obvious example, the Catholic Church sets limits on the autonomy of the state with regard to various ‘moral’ issues.

[49] Cf. Kagan and Skolnick 1993.

[50] ‘Close-knit group’ differs from ‘SASF’ in that, as Ellickson uses the term, he seems to contemplate a non-continuous variable (groups being either close-knit or not). There seems to be nothing in his approach, however, that would commit him so such a position.

[51] Cf. Moore 1973; Collier 1976.

[52] Compare Macaulay 1963.

[53] This example derives from P. Oden’s research concerning the social working of Dutch affirmative action measures in connection with the appointment of heads of primary schools (Oden 1993). Compare E. Abel 1981.

[54] Resistance to ‘external’ law (Kidder 1979) can extend over many years. A legal system prepared to invest in vigorous, proactive enforcement can temporarily achieve a fair measure of outward conformity, so that the appearance of social change is achieved. When the pressure falls away, however, it becomes clear that old behavioral rules are still very much alive. In recent years, news from Eastern Europe has regularly confirmed this possibility (e.g. in relation to racism and antisemitism). Cf. also Massell 1968.

[55] See Homans 1951.

[56] See e.g. Freidson 1975; Sudnow 1976; Anspach 1993.

[57] See Hart 1961.

[58] The ‘legitimacy’ of an external rule depends on the sovereignty rules prevalent on the shop floor in question. ‘Legitimacy’ can be defined as the extent to which “communication about the existence of a law ... may change the moral evaluation ... of a specific item of conduct” (Galanter 1981: 12). When a legal rule is regarded by the sovereignty rules of a SASF as an acceptable reason for the required conduct, the rule is ‘legitimate’ as far as the SASF is concerned, although its choice-of-rule rules may not provide for the rule’s local enforcement.

[59] Ellickson’s (1991) discussion of the variety of sorts of primary and secondary rules in a ‘close-knit group’ is very similar to the discussion here. Compare also Honoré 1987. Verkruyten et al. (1994) report an interesting social-psychological study of ‘rules for breaking formal rules’ (shared attitudes on the circumstances in which it is considered justifiable to evade taxes or run a red light). While from the point of view of the argument being made here it does seem likely that social rules lie behind the expressed attitudes, the authors (despite the promise of their article’s title) do not explore the possibility.

[60] ‘Encounter’ perhaps suggests that regulation by other SASFs always antedates external ‘legal’ regulation of shop-floor behavior. That is of course not the case. State law can in various ways lead to the emergence of local-level regulation, and even to the emergence of new SASFs that engage in regulation. The relationship between the state and other SASFs, and their respective regulatory activities, is in historical perspective dynamic and interactive (cf. Collier 1976). This additional layer of chronological complexity is not important for my present argument, which can be more simply expressed in anachronistic terms.

[61] For the sake of simplicity, I do not distinguish euthanasia proper (killing a person at his request) from assistance with suicide. For most purposes the distinction is not important in Dutch law.

[62] The normative situation is actually far more complicated than this since the rule requiring beneficence conflicts to some extent with other rules requiring respect for the autonomy of the patient and the saving of human life. All these rules exist in legal, general moral and professional forms, albeit with differences of relative weight: the rule requiring respect for autonomy, for example, has at least until recently probably been less weighty as a professional than as a legal and general social rule. The relative weight attached to the various rules also differs from one medical and legal culture to another, which may partly account for the fact that in the Netherlands euthanasia has been widely supported by doctors in terms of the rule requiring alleviation of suffering, whereas in the United States it has been generally [page 31] resisted by the medical profession, and advocated by others, in the name of patients’ rights.

[63] The element of moral obligation is important since it is well established that most Dutch doctors perform euthanasia only with great reluctance, and that a patient who wants it must overcome considerable resistance. See Griffiths, Bood, and Weyers 1998: 247-248; see Chabot 2001 for more recent information in this regard.

[64] See section 6 of this article on the role of the medical profession in the process of legal change.

[65] I believe that Hart’s categories of secondary rules require three additions: ‘choice-of-rule’ and ‘sovereignty’ rules (discussed in the text) and ‘constitutional’ rules. As to the latter, it seems important to note that SASFs often - probably always - have rules limiting the legitimate regulatory authority of the group. The bite of such constitutional rules is immediately felt when someone tries to get a group to address a subject outside the scope of what is taken to be its raison d’être (for example, ‘political issues’ such as the War in Vietnam in a faculty meeting, to invoke an example from my own experience). In the past, smoking was often considered an issue that ‘ought not to be addressed’ in many fora (clubs, committees, etc.). I remember a rather nasty clash between Americans and Europeans over just this ‘constitutional’ issue at an international conference in the early 1980s. Regulation of personal life-style when not engaged in group activities - something that until quite recently many groups felt entitled to do - is now generally deemed outside the pale [page 32] of group authority. In short, there is nothing static about the ‘constitutional law’ of SASFs.

[66] Internal hospital regulations and employment contracts, for example, often require such reporting. Compare Freidson 1975 on the unwillingness of doctors to report on each other.

An instructive case involving a psychiatrist who committed suicide after his suspension from practice in a psychiatric hospital became public, was reported in the press a few years ago (Trouw, 1 February 1996). After a colleague brought the psychiatrist’s “seriously disturbed therapeutic relationship” with a patient to the attention of the hospital authorities, the latter sought the assistance of the Medical Inspectorate. When the Inspectorate confirmed the judgment of the hospital authorities, they suspended the psychiatrist and made this action known to the psychiatric profession at large. After the suicide, the hospital authorities were severely criticized by psychiatrists elsewhere for not having dealt with the case in the proper way, “behind closed doors”; they blamed the suicide on the “indiscrete” actions of the hospital authorities.

Inappropriate sexual behavior between teachers and pupils, and between priests and those for whom they are responsible (adult and minor), is currently a matter of public concern in the Netherlands and elsewhere. Effective control over this sort of behavior is particularly difficult because of the rule requiring that potential scandal be dealt with privately and internally. In this case legal authorities seem to be bent on breaking the autonomy of SASFs such as schools and churches, and as far as one can tell they are succeeding.

[67] The position many Dutch doctors took early on in the development of Dutch euthanasia law ‑ that euthanasia decisions pertain to the doctor-patient relationship, a moral domain the state has no business trying to regulate ‑ was often couched in terms reminiscent of the position virtually the entire Dutch medical profession took during the most dramatic sovereignty struggle of its existence: its collective resistance to the efforts by the German occupation authorities to secure effective control of the doctor-patient relationship during the Second World War. On several occasions, anticipating efforts to submit the profession to political control, the doctors’ resistance organization was able to mount highly successful counter-actions. In one such action, more than 6200 doctors (over 95% of all Dutch doctors at the time) resigned their licenses to practice medicine as a means to frustrate a [page 34] particularly threatening effort to control them; they were able to force the Germans into a compromise. What was at stake in the doctors’ resistance was not opposition to the German occupation as such, but a specific professional objection to political control of the doctor-patient relationship. Time and again the doctors’ resistance organization emphasized that the relationship of doctor and patient is one of trust and confidentiality and that efforts to control the relationship (e.g. in connection with venereal disease) or to violate its confidentiality (e.g. by requiring reporting of gunshot wounds or of Jewish patients) were intolerable. On the doctors’ resistance see De Vries 1949: ch. 5; De Jong 1974, 1975, 1976.

The same sovereignty rule seems to have been an important basis for the systematic frustration by Dutch doctors of the legal prohibition on abortion in the period before abortion law reform: see Outshoorn 1986: 115-123.

[68] Such ‘sovereignty’ rules are often expressed in terms of a sphere of ‘privacy’ into which the state should not enter. In Aubert’s (1966) study of the effectiveness of Norwegian legislation concerning the rights of household personnel, for example, social rules excluded external labor law from the domestic scene.

[69] See Adams 2001: 46.

[70] See many of the references in note 46.

[71] Compare Taale and Griffiths (1995) on Ecuadorian forest-protection legislation. In an ironic reverse of the classic picture of ineffectiveness due to the resistance of strong local SASFs, in the Amazonian situation the state has little enforcement capacity of its own and sought to enlist the support of local SASFs. This strategy seemed unlikely to succeed because, as is typical of ‘frontier’ situations (cf. Doyle 1978), the local SASFs were so feeble they probably would not be capable of enforcing forestry regulations even if inclined to do so.

Similarly situated groups may exhibit considerable differences in their ability to enforce external rules. Horwitz (1990: 234-236) argues that the considerable effectiveness in the homosexual community of legal measures to prevent the spread of HIV and the low effectiveness of the same measures among drug users is a result of the strong social organization of the former group and the weak social organization of the latter.

[72] See Kagan and Skolnick 1993. Compare the examples of the queuing rule at London bus stops and the New York City clean-up-after-your-dog ordinance, described in note 74.

[73] See notes 58 and 65 on 'sovereignty' and 'constitutional' rules limiting 'legitimate' authority.

[74] Compare the alleged effectiveness in New York City of a local ordinance requiring dog owners to clean up after their dogs. In the 1970s, when I lived in New York, the absence of clear rules permitted a reign of terror by dog owners (when I once remonstrated with the owner of two gigantic dogs, doing their thing on one of the few bits of grass near our apartment complex where children could play - while there was a dog run at about 100 yards distance - I got the following response, whose tone was as aggressive as its content: “You fucking homosexual communist! Go back to Russia where you come from!”). The ordinance is said to have led to a radical change in the moral balance of power between dog owners and others (see F. Kotterer, ‘Ooooh, shit!’ Het Parool, 14 September 1996).

A few years ago, London Transport announced the abolition of the legal requirement of queuing at bus stops, which it regarded as having fallen into disuse, no one being able to remember “the last time a prosecution had been brought for failure to respect a line”. London Transport does not expect any effect on the disciplined behavior of local users of the bus service: “We’re still a fairly well-mannered society.” It will be interesting to follow this experiment in social control to see whether informal social control indeed no longer needs external specification of the rule of behavior (see International Herald Tribune, 31 December 1994/1 January 1995). Compare Reisman (1999) on the non-legal rules governing queuing.

[75] On ‘symbolic’ effects see note 15 above.

[76] Compare Ellickson (1991: 258-264) (lack of support in universities for rules concerning the copying of materials for classroom use).

[77] See Black 1973; Blankenburg 1980, 1995.

[78] A boarding school I attended almost half a century ago had the following such rule, which it included in the instruction leaflet sent to new students: ‘borrowing without permission is considered stealing’.

[79] Some first steps are taken in Schwartz 1954 (social control in Israeli agricultural communities); Macaulay 1963 (social control among businessmen); Homans 1951 (social control in the workplace); Todd 1978; Homans 1951; Griffiths 1984c; Ellickson 1991 (social control in village and similar communities). Verkruisen’s (1993) careful and detailed analysis of the ways patients invoke rules in their dealings with doctors is limited to the context of (potential) disputes, but is both theoretically and empirically the most sophisticated treatment to date.

[80] See e.g. Aubert 1966.

[81] There is great variation in the degree of tolerance that the state exhibits toward SASFs. Some SASFs enjoy a measure of toleration or even of respect and are afforded a wide latitude for private self-regulation (e.g. the household - Aubert 1966; Black 1980), sometimes in the form of various ‘immunities’ in legal doctrine itself (e.g. schools, the medical profession, established churches); others receive an indifferent or even an actively hostile reception at the hands of the law (e.g. new religious sects). Explanation for these differences (cf. Black 1976) is beyond the ambition of this article.

[82] See Verkruisen 1993: 7, 10 n9.

[83] See Snell 1979.

[84] See Schwartz 1954; compare Ellickson 1991.

[85] See Griffiths 1984c; compare Jacobs 1961.

[86] See generally Griffiths, Bood, and Weyers 1998.

[87] There are important variations in the social organization of family practice, which are reflected in differences in the degree of mutual scrutiny. In Amsterdam, for example, where euthanasia (at least in the early years) was apparently most frequent and the social organization of medical practice surrounding euthanasia already relatively tight, the medical association and the Dutch Government recently established a special service to assist doctors in euthanasia cases (see Klijn 2002: 164-165); after a positive evaluation, the program is now being extended to the whole country. One effect will presumably be an increase in legal intelligence. The growth over recent years of joint practices, primary health care centers, and the like will have a similar effect.

[88] See, for the situation of medical practice, Freidson 1975; Sudnow 1976; Anspach 1993.

[89] Unfortunately, as far as I have been able to discover, there is no good analytic literature on this well-known social phenomenon.

[90] It has frequently been observed (e.g. Aubert 1966) that the language used in legislation is a substantial barrier to communication with non-official mobilizers.

[91] Hawkins’ description of water-pollution officials in England finds echoes throughout the literature:

Knowledge of the law relating to control of water pollution, beyond a broad conception of the pollution offences ... is regarded as unimportant and is claimed by only a small minority of field men, because the job is done ‘by experience’ and the application of rules-of-thumb - not ‘by the book’. As long as they know what kinds of acts or events might be prosecutable, they take the view that the arcane business of the law is best left to senior staff. (Hawkins 1984: 42)

[92] See Vezzoni and Griffiths 2001.

[93] The amount and the quality of the legal information lawyers transmit has been the object of some study, from which it can be concluded that what they tell their clients about the ‘law’ sometimes differs significantly from the applicable legal rules (see e.g. Macaulay 1979; Griffiths 1986b). Macaulay shows that information costs in the context of the economics of law practice are important factors in this regard; this is confirmed by our research on the social working of advance directives: what Dutch clients hear from their lawyers concerning the drafting of an advance directive is often legally incorrect and otherwise uninformed precisely because it would be too expensive for them to acquire accurate knowledge. But both Macaulay and Griffiths make clear that both professional rules and the behavioral expectations of other [page 48] actors in the social fields in which lawyers are active also strongly influence what lawyers tell their clients about ‘the law’.

Compare Edelman, Fuller and Mara-Drita 2001, and Edelman, Erlanger and Lande 1993, for empirical descriptions of the process of “managerialization of law”: the transformation by personnel officers of large organizations of the legal concepts of anti-discrimination law into concepts of ‘good management’.

[94] See Bohannan 1965; Collier 1976.

[95] In practice, of course, the chronological order may well be the other way around, an actor only coming to know about a legal rule when he already sees the matter involved as a ‘legal’ one. The two requisites taken together ‑ knowing a rule and regarding it as relevant to a specific situation ‑ have sometimes been called ‘thematization’ (Blankenburg 1995), this being conceived as the first step in the process of ‘mobilization’ of a legal institution.

[96] See e.g. Aubert 1966.

[97] See e.g. Macaulay 1979; Galanter 1974.

[98] For a detailed description of these rules see generally Griffiths, Bood, and Weyers 1998.

[99] Id. at p. 57. Although at the time Hilhorst wrote euthanasia was still forbidden, and knowledge of how to do it in a legal way therefore - at least formally speaking - not yet relevant, the rules distinguishing ‘euthanasia’ from other medical practices that lead to the death of the patient (abstention and pain relief) were obviously of great (legal) importance. These, too, were unknown to doctors.

[100] Informal communication. Discussions with members of several Regional Assessment Committees confirm this judgment. Doctors’ inability to reproduce legal rules with any accuracy does not, of course, preclude more subtle and diffuse ways in which the flood of public information on the subject may be influencing behavior. The conceptual/terminological and normative change that has taken place in Dutch society - and among Dutch doctors - is unmistakable. When people begin to think and talk about behavior in different terms they probably begin to behave differently too. When the term ‘euthanasia’ became current as the name of a legitimate sort of termination of life on request, it became easier to discuss it in public and, probably, to carry it out in a self-conscious and careful way. It seems safe to assume that recognition of the fact they are engaged in legally regulated behavior encourages doctors to seek out legal information when confronted with a concrete case.

[101] Griffiths, Bood, and Weyers 1998: 248-251.

[102] Id., p. 216 and n. 50.

[103] See Griffiths, Bood, and Weyers 1998, and in particular Griffiths 1999a, for the following analysis.

[104] Cf. Jackson 1996: 93-98.

[105] See Griffiths 1999a for a fuller discussion. Different conceptions of causation, to give another example, lead to differing interpretations of the act/omission distinction and of the ‘cause’ of a patient’s death. It appears from a recent incident, for example, that Dutch obstetricians generally regard a death resulting from a medical mistake during childbirth (in the case that brought the problem to light: unskillful use of forceps) as being ‘caused’ by the medical condition that gave rise to the need to [page 52] intervene, and not to the unskillful intervention, with the important consequence for legal control that the death can be reported as ‘natural’ and therefore attract no legal attention. The Medical Inspector publicly announced his disagreement with this point of view and began disciplinary proceedings to drive the point home (the case apparently ended inconclusively). See Trouw 27 March 2000.

[106] See Griffiths, Bood, and Weyers 1998: 162 and elsewhere; Otlowski, 1997: 170-186 (for the common-law countries).

[107] Statistical evidence to the same effect has been collected by D. van Tol in the context of the research program described in note 5; it will be presented and analyzed in his forthcoming dissertation.

[108] It is important to realize, in this connection, that if the doctors had called what they did ‘termination of life without a request’ it would probably have been deemed justifiable under current Dutch law; it is unlikely that there would have been a prosecution, and if there had been, the doctors would probably have been acquitted (see Griffiths, Bood and Weyers 1998: 123ff.). This legal information was known to the doctors involved.

[109] This conclusion is argued in detail on the basis of Dutch empirical data in Griffiths, Bood, and Weyers 1998: 254ff., 269ff.

[110] Law on Contracts for Medical Treatment, effective 1 April 1995 (artt. 7:447 ff. of the Dutch Civil Code).

[111] These reactions were probably disingenuous, since the icon clearly reads ‘Silence’. It does, however, illustrate how problems of interpretation can frustrate attempts at mobilization.

[112] Compare Macaulay 1979: 159.

[113] The sociological principle underlying such cases is the ‘relational distance’ thesis: differentiation in social control increases with increasing relational distance (Griffiths 1984a). The mobilization of such a highly differentiated form of social control as law, and still more so the mobilization of legal specialists and institutions, generally occurs only between parties who share no social relationships that are important to them. See generally Black (1976) on the sociology of mobilization.

[114] Cf. Ellickson 1991.

[115] Cf. e.g. Galanter 1974; Macaulay 1979.

[116] Compare Hyde 1983; Merry 1990; Scheingold 1974.

[117] I owe the information on this case to my colleague Rob Schwitters. See S. Ottesen, Ma jeg do i smerte? Oslo: Aschehoug, 2002.

[118] See Taale and Griffiths 1995; Ellickson 1991; Macaulay 1963; Moore 1973.

[119] The effectiveness of the secondary rule against reporting misbehavior of one’s colleagues to external authorities probably explains the fact that in the Netherlands, with an extensive and quite open euthanasia practice, there have over a period of more than 20 years been almost no complaints to the authorities by one medical professional against another. Compare Magnusson’s (2002) findings for Australia and the United States on the absence of complaints by one doctor against another for having practiced euthanasia.

[120] See Black 1973; cf. also Aubert 1966; Mayhew 1968: 273.

[121] See e.g. Pressman and Wildavsky 1973; Hawkins 1984.

[122] H.L.A. Hart (1961) refers to these as the ‘secondary rules’ of a legal system: the rules that obtain among legal professionals determine what counts as a ‘legal’ rule, and are thereby constitutive of ‘law’ as a social phenomenon. In section 4 we have already seen that such secondary rules are not limited to ‘lawyers’ but are an important part of the normative property of all SASFs.

[123] See e.g. Kagan 1978. All of this applies mutatis mutandis to courts as well. It is the secondary rules of the court and the judiciary as a whole, and of the local and national legal profession, far more than the individual attitudes and cognitive processes that have played such a predominant role in the literature on judicial decision-making, that largely determine the process of legal reasoning. See Griffiths 1986c.

[124] See Hawkins 1984; Kagan 1989.

[125] See e.g. Hawkins 1984.

[126] See Griffiths, Bood, and Weyers 1998.

[127] As we will see in section 6, these professional guidelines amounted to professional self-regulation, anticipating the development of legal rules.

[128] I have this quotation on the authority of one of his students, Marc Galanter, who heard it in the classroom.

[129] See Griffiths, Bood, and Weyers 1998: ch. 6.3.

[130] Cf. Griffiths 1985.

[131] There are a variety of problems with this approach, of course. Two of the more fundamental concern the notion of (non-normative) ‘preferences’ concealed in the idea of welfare-maximization, and the assumption that a collective good such as a rule can be produced by ‘rational actors’ pursuing their self-interest. Cf. Griffiths 1995b.

[132] The importance of the difference between the two sorts of explanation is familiar from evolutionary biology. That a characteristic is functional for a species explains why those members of the species who possess the characteristic are reproductively [page 68] more successful than their fellows who lack it: the ‘fittest’ survive. What functionality cannot explain is how the characteristic got acquired in the first place.

[133] While the term ‘self-regulation’ is common in the (Dutch) socio-legal literature, it usually refers to one or another technique of governance (conceived in a top-down way): decentralization, covenants, etc.

[134] See e.g. Edelman n.d., 2002; Edelman, Uggen and Erlanger 1999.

[135] I have rather simplified the facts here. See Griffiths, Bood, and Weyers 1998: 51 ff. for a fuller account.

[136] See Griffiths, Bood, and Weyers 1998; Weyers 2001.

[137] See Griffiths 2000. The last, legislative stage has yet to occur. Less clear-cut processes have also taken place with respect to patients in a ‘persistent vegetative state’, late-term abortion and abstaining from artificial feeding and hydration of patients with advanced dementia. See Griffiths, Bood, and Weyers 1998: 126-131, 134-137.

[138] I limit the discussion here to primary rules. The story becomes more complicated but not significantly different if secondary rules are included as well.

[139] I have changed a few minor facts in the following account to protect the anonymity of the respondent and the ICU.

[140] This local rule is the same as the legal rule on the subject. Whether the doctors involved knew this is not clear.

[141] As Sudnow (1976) has described, doctors and nurses do not like to deal with dead patients.

[142] Compare Fallers (1969) for the application to a situation of unwritten, customary law of Levi’s (1949) idea that gradual, piecemeal legal change in the context of a stable set of rules characteristically takes because the unchanging terms in the rules are subject to a ‘moving system of classification’.

[143] Cf. Eisenberg (1976) for the argument that the formal legislative process often similarly consists of the mobilization of existing rules.

[144] See Jutras 2001; compare Macdonald 2002.

[145] Although he did not succeed in convincing me of the necessity of this conclusion at the time, some nagging questions about the relationship between the social working and the social genesis of legal rules that my colleague Rob Schwitters planted in my mind several years ago, together with the theoretical stimulation of ongoing research, have finally produced results.

[146] Since the precise contours of what constitutes ‘legislation’ are not important for the argument, no attention has been directed to that question.

[147] That social behavior is often incomprehensible unless one pays attention to the rules being followed has been emphasized by writers from a variety of perspectives. See e.g. Ellickson 1991; Sunstein 1996; Reisman 1999; Jutras 2001; Macdonald 2002.

[148] Quoted by F. von Benda-Beckmann 1989: 129.

THE SOCIAL WORKING

OF LEGAL RULES

John Griffiths